ABSTRACT

The objective of this study is to process the soil images to generate a digital soil classification
system for rural farmers at low cost. Soil texture is the main factor to be considered before doing
cultivation. It affects the crop selection and regulates the water transmission property. The
conventional hydrometer method determines the percentage of sand, silt, and clay present in a
soil sample. This method is very cost and time-consuming process. Soil images are processed
through the different stages like pre-processing of soil images for image enhancement,
extracting the region of interest for segmentation and the texture analysis for feature vector. The
feature vector is calculated from the Hue, Saturation, and Value (HSV) histogram, color
moments, color auto Correlogram, Gabor wavelets, and discrete wavelet transform. Finally,
Support Vector Machine classifier is used to classify the soil images using linear kernel. The
proposed method gives an average of 91.37% accuracy for all the soil samples and the result is
nearly the same with the any country of Agriculture soil classification.

keywords:

Soil classification, SVM (support vector machine),soil texture

CHAPTER 1

INTRODUCTION

The physical and chemical properties of the soil always play an important role in precision and
smart farming. Soil texture is the main physical property of the soil and plays an important role
in farming [39]. It defines the fraction of sand, silt, and clay present in a soil sample. Other
related agriculture cultivating properties such as water content of the soil, plant growth, and
crop selection relate to these three end members. Most of the rural farmers are unknown
about the texture and characteristics of the soil. Sandy soil is low in the water holding
capacity and organic matter. The water holding capacity of silt soil is more as compared to
sandy soil [25]. Clay soil shows the high water holding capacity, high plasticity and
thickness whereas sandy soils are conspicuous by the absence of these properties. The water
holding capacity of sandy loam is low as compared to loam and silty clay soil [2]. It
indirectly affects the plant growth, the weight of the plant and the percentage of chlorophyll. For
example, chlorophyll contents of leaves are dropped in sandy loam as com- pared to silty clay
soil [2].
In a number of engineering problems there is a necessity to classify contiguous intervals
(segments) of series data (signals). Series data has an additional index variable (distance or time)
associated with each data value. Standard classification algorithms in these situations are often
inadequate due to the additional contiguity constraint. Examples from the following domains can
be mentioned: classification of sub-soil layers using Cone Penetration Testing [2] [7], well-log
analysis in petroleum engineering [9], palaeoecology [4], etc. In these cases measurements are
taken from a vertical bore or with a test apparatus which is pushed down the earth and it is
required that the stratigraphical information is preserved in the classification. The problem is
solved in two phases: firstly, a segmentation algorithm is used to cluster contiguous blocks of
instances and secondly, these segments are classified by domain experts. We investigated the
problem with a specific interest of automating classification of soil layers from measured data.
In civil engineering it is a prerequisite to know the soil classes up to some depths prior to any
construction. The direct method to identify the soil classes by drilling and testing soil samples is
very expensive. A cheaper alternative is the so-called Cone Penetration Testing (CPT) which is
one of the most popular soil investigation methods [2]. In CPT, a metallic cone is pushed into
the soil and an indication of the in-situ soil strength is obtained by measuring the force needed to
let it advance at a constant rate. A CPT recording is a quasi-continuous picture of the subsurface
at the test location. It contains the vertical variations of the mechanical characteristics of the
subsoil. These variations in turn indicate variations in geological layers and their properties.
During a test, two primary signals are recorded: 1) the cone tip resistance stress (qc), 2) the
frictional stress (f1) which is used to derive the more widely used friction ratio Rf = f *100Iq.
Additionally, information is available from borehole drilling in the proximity of CPTs typically
with the frequency of I borehole for 10 CPTs. Observing the variations of q, and Rf (Fig.1) and
using the nearby borehole information, an expert firstly segments the logs i.e., finds boundaries
of layers (class boundaries), and secondly, using the domain knowledge assigns a soil class C, to
each segment (where i =,2, ...,I and I=number of classes). In practice a manual segmentation and
classification procedure is followed. This procedure requires expertise, and is expensive, time
consuming, subjective and not completely reproducible. The challenge is to automate this
procedure. In order to achieve this anew algorithm called CONCC (CONstraint Clustering and
Classification) was developed. It can be used in automatic classification with the constraint of
contiguity and includes the following two steps: - Segmentation: to find J segments of data from
a single series data(e.g. CPT); - Classification: to build a classifier to assign classes to these
segments; it is built using measured data and extracted features from segments from a number of
series data from a region and trained with classes labelled by experts. Segmentation of series
data has been addressed in [1], and is briefly reported here. This presents classification of the
found segments using Support Vector Machines (SVM). Application of ML in geostatistical
problems is quite limited, some applications are reported in [1] [8] [10].

Conventionally, the hydrometer method calculates the fraction of sand, silt, and clay of soil.
Then classifies the soil. All these methods are time and labour consuming processes. These are
not appropriate approaches for rural farmers. In the field study, it is found that rural farmers do
not have any knowledge about the soil texture. They are doing their farming without any proper
soil testing and are unaware about the selection of soil and seed. It indirectly affects the overall
growth of the plant. To over- come this problem, researches are done to develop the computer
system and it efficiently classifies the soil texture using digital image processing.

Most of the work of soil texture classification is in three classes or multi-class. Among the three
classifications, Zhang et al. [4] classified sand, silt, and clay soil using wavelet trans- form and
maximum likelihood method. They incorporated fisher’s linear discrimination analysis (FLDA)
for feature dimension reduction and optimization.

CHAPTER 2

LITERATURE SURVEY

This section deals with many methods that previously discussed that provide valuable
information regarding different soils.

Sun et al. [1] for- warded a wavelet frame based method to classify the three classes of soil
texture and achieved 91% accuracy for the classification. They achieved the same accuracy
using Gabor wavelets [1]. Bhattacharya and Solomatine [8] introduced the decision tree,
artificial neural network and support vector machine classifier to classify the soil texture on the
segmented soil samples and achieved 90.7% accuracy for support vector machine.

Zaho et al. [26] presented a model to predict the soil texture of high-resolution soil images.
They applied artificial neural network to predict the sand and clay con- tents of the soil in the
collected soil maps. The neural net- work model was trained by the Levenberg-Marquardt
optimization algorithm and Resilient back propagation algorithm but achieved more accuracy in
the Levenberg- Marquardt optimization algorithm. The accuracy for Clay was 88% whereas the
accuracy for Sand classified the clay, loam and sandy soil of the Yangtze River, Southwest
China. They compared the performance of different machine learning approaches for soil
classification such as support vector machine, artificial neural network, and classification tree.
The support vector machine with the polynomial kernel gave a better result as compared to the
other two one.
Among the multi-classification, Zhang et al. [5] introduced a discrete wavelet transform and
linear discriminate analysis method for soil texture classification. They mixed the hyper spectral
signals of sand, silt, and clay soil and classified using the maximum likelihood classifier. Later,
they applied the Hidden Markov model for soil texture classification and achieved more
accuracy as compared to maximum likelihood classifier [6].

Bhattacharya and Solomatine [8] applied decision tree, artificial neural network and support
vector machine classifier to classify the soil texture from the segmented soil sample. They
achieved an average of 70.4% accu- racy for the classification and found that the seven support
vector machine classifier gave the best result for silty sand and clay sand. Artificial neural
network and decision tree model gave the best result for humus clay, peat, and the clay peat soil
respectively.

Shenbagavalli et al. [7] presented a soil texture classification method using mask convolution.
They computed mean, abs mean, standard deviation, skewness, and kurtosis of the soil image
and prepared as a feature vector. Chung et al. [9,36] introduced a soil texture classification
model using RGB color. The RGB values are regressed with the actual laboratory values and
got the coefficient of best fit as 0.96.

Vibhute et al. [22] applied the radial base kernel of the SVM classifier on hyperspectral remote
sensing soil data and classified five types of soil. The average accuracy of the model was
71.18% with the Kappa = 0.57. Breul and Gourves [3] presented a method of soil
characterization using the third order moment of the input image. They recorded soil images
using an endoscope.

It is come to know that soil texture property is the most important property for plant growth.
During the literature survey, it is found that the researchers did valuable works for soil texture
classification. However, most of the works are not in the aspect of a rural farmer as they
collected their soil images using sensor or used hyper spectral soil images [4,8,22]. Sometimes
the researchers collected the soil samples using a digital camera [9]. In this paper, 50 soil
samples are photographed using an Android mobile device with 13 MP cameras. Capturing soil
image using a Smartphone is the main challenging task for soil classification because the natural
light can directly affect the quality of the soil image. Measuring all these factors, we developed
our system in such a way that a rural farmer can easily capture a soil sample and use the system
for soil classification.
CHAPTER 3

INTRODUCTION TO IMAGE PROCESSING

GENERAL

The term digital image refers to processing of a two dimensional picture by a digital computer. In a
broader context, it implies digital processing of any two dimensional data. A digital image is an array of
real or complex numbers represented by a finite number of bits. An image given in the form of a
transparency, slide, photograph or an X-ray is first digitized and stored as a matrix of binary digits in
computer memory. This digitized image can then be processed and/or displayed on a high-resolution
television monitor. For display, the image is stored in a rapid-access buffer memory, which refreshes the
monitor at a rate of 25 frames per second to produce a visually continuous display.

3.1.1 THE IMAGE PROCESSING SYSTEM

Digitizer Mass Storage

Image Digital Operator

Processor Computer Console

Hard Copy
Display Device

FIG 3.1 Block Diagram For Image Processing System

3.2 DIGITIZER:

A digitizer converts an image into a numerical representation suitable for input into a digital computer.
Some common digitizers are

Microdensitometer

Flying spot scanner

Image dissector
Videocon camera

Photosensitive solid- state arrays.

3.3 IMAGE PROCESSOR:

An image processor does the functions of image acquisition, storage, preprocessing, segmentation,
representation, recognition and interpretation and finally displays or records the resulting image. The
following block diagram gives the fundamental sequence involved in an image processing system.

Problem Image Representation &

Image Segmentation Representation &
Acquisition Segmentation Description
Domain Acquisition n Description
n

Knowledge Result
Preprocessing Knowledge Recognition &
Preprocessing Recognition &
Base interpretation
Base interpretation

FIG 3.2 Block Diagram Of Fundamental Sequence Involved In An Image Processing

System

As detailed in the diagram, the first step in the process is image acquisition by an imaging sensor in
conjunction with a digitizer to digitize the image. The next step is the preprocessing step where the image
is improved being fed as an input to the other processes. Preprocessing typically deals with enhancing,
removing noise, isolating regions, etc. Segmentation partitions an image into its constituent parts or
objects. The output of segmentation is usually raw pixel data, which consists of either the boundary of the
region or the pixels in the region themselves. Representation is the process of transforming the raw pixel
data into a form useful for subsequent processing by the computer. Description deals with extracting
features that are basic in differentiating one class of objects from another. Recognition assigns a label to
an object based on the information provided by its descriptors. Interpretation involves assigning meaning
to an ensemble of recognized objects. The knowledge about a problem domain is incorporated into the
knowledge base. The knowledge base guides the operation of each processing module and also controls
the interaction between the modules. Not all modules need be necessarily present for a specific function.
The composition of the image processing system depends on its application. The frame rate of the image
processor is normally around 25 frames per second.

DIGITAL COMPUTER:

Mathematical processing of the digitized image such as convolution, averaging, addition, subtraction, etc.
are done by the computer.

MASS STORAGE:

The secondary storage devices normally used are floppy disks, CD ROMs etc.

HARD COPY DEVICE:

The hard copy device is used to produce a permanent copy of the image and for the storage of the
software involved.

OPERATOR CONSOLE:

The operator console consists of equipment and arrangements for verification of intermediate results and
for alterations in the software as and when require. The operator is also capable of checking for any
resulting errors and for the entry of requisite data.

IMAGE PROCESSING FUNDAMENTAL:

Digital image processing refers processing of the image in digital form. Modern cameras may directly
take the image in digital form but generally images are originated in optical form. They are captured by
video cameras and digitalized. The digitalization process includes sampling, quantization. Then these
images are processed by the five fundamental processes, at least any one of them, not necessarily all of
them.

3.5 IMAGE PROCESSING TECHNIQUES:

This section gives various image processing techniques.

Image Enhancement

Image Restoration

IP
Image Analysis
 Image Compression

Image Synthesis
Fig 3.3: Image Processing Techniques

IMAGE ENHANCEMENT:

Image enhancement operations improve the qualities of an image like improving the image’s contrast and
brightness characteristics, reducing its noise content, or sharpen the details. This just enhances the image
and reveals the same information in more understandable image. It does not add any information to it.

IMAGE RESTORATION:

Image restoration like enhancement improves the qualities of image but all the operations are mainly
based on known, measured, or degradations of the original image. Image restorations are used to restore
images with problems such as geometric distortion, improper focus, repetitive noise, and camera motion.
It is used to correct images for known degradations.

IMAGE ANALYSIS:

Image analysis operations produce numerical or graphical information based on characteristics of the
original image. They break into objects and then classify them. They depend on the image statistics.
Common operations are extraction and description of scene and image features, automated measurements,
and object classification. Image analyze are mainly used in machine vision applications.

IMAGE COMPRESSION:

Image compression and decompression reduce the data content necessary to describe the image. Most of
the images contain lot of redundant information, compression removes all the redundancies. Because of
the compression the size is reduced, so efficiently stored or transported. The compressed image is
decompressed when displayed. Lossless compression preserves the exact data in the original image, but
Lossy compression does not represent the original image but provide excellent compression.

IMAGE SYNTHESIS:

Image synthesis operations create images from other images or non-image data. Image synthesis
operations generally create images that are either physically impossible or impractical to acquire.

3.6 APPLICATIONS OF DIGITAL IMAGE PROCESSING:

Digital image processing has a broad spectrum of applications, such as remote sensing via satellites and
other spacecrafts, image transmission and storage for business applications, medical processing, radar,
sonar and acoustic image processing, robotics and automated inspection of industrial parts.

MEDICAL APPLICATIONS:

In medical applications, one is concerned with processing of chest X-rays, cineangiograms, projection
images of transaxial tomography and other medical images that occur in radiology, nuclear magnetic
resonance (NMR) and ultrasonic scanning. These images may be used for patient screening and
monitoring or for detection of tumors’ or other disease in patients.

SATELLITE IMAGING:

Images acquired by satellites are useful in tracking of earth resources; geographical mapping; prediction
of agricultural crops, urban growth and weather; flood and fire control; and many other environmental
applications. Space image applications include recognition and analysis of objects contained in image
obtained from deep space-probe missions.

COMMUNICATION:

Image transmission and storage applications occur in broadcast television, teleconferencing, and
transmission of facsimile images for office automation, communication of computer networks, closed-
circuit television based security monitoring systems and in military communications.

RADAR IMAGING SYSTEMS:

Radar and sonar images are used for detection and recognition of various types of targets or in guidance
and maneuvering of aircraft or missile systems.

DOCUMENT PROCESSING:

It is used in scanning, and transmission for converting paper documents to a digital image form,
compressing the image, and storing it on magnetic tape. It is also used in document reading for
automatically detecting and recognizing printed characteristics.

DEFENSE/INTELLIGENCE:

It is used in reconnaissance photo-interpretation for automatic interpretation of earth satellite imagery to

look for sensitive targets or military threats and target acquisition and guidance for recognizing and
tracking targets in real-time smart-bomb and missile-guidance systems.

3.7 Introduction to Digital Image Processing

Image processing is any form of signal processing for which the input is an image, such as
a photograph or video frame; the output of image processing may be either an image or, a set of
characteristics or parameters related to the image. Most image-processing techniques involve treating the
image as a two-dimensional signal and applying standard signal-processing techniques to it. Image
processing usually refers to digital image processing, but optical and analog image processing are also
possible.

Gray scale image:

A grayscale image is a function I (xylem) of the two spatial coordinates of the image plane.

I(x, y) is the intensity of the image at the point (x, y) on the image plane.

I (xylem) takes non-negative values assume the image is bounded by a rectangle [0, a] ´[0, b]I: [0, a] ´ [0,
b] ® [0, info)

Color image:

It can be represented by three functions, R (xylem) for red, G (xylem) for green and B (xylem)
for blue.

An image may be continuous with respect to the x and y coordinates and also in
amplitude. Converting such an image to digital form requires that the coordinates as well as the amplitude
to be digitized. Digitizing the coordinate’s values is called sampling. Digitizing the amplitude values is
called quantization.

Coordinate convention:

The result of sampling and quantization is a matrix of real numbers. We use two principal ways to
represent digital images. Assume that an image f(x, y) is sampled so that the resulting image has M rows
and N columns. We say that the image is of size M X N. The values of the coordinates (xylem) are
discrete quantities. For notational clarity and convenience, we use integer values for these discrete
coordinates. In many image processing books, the image origin is defined to be at (xylem)=(0,0).The next
coordinate values along the first row of the image are (xylem)=(0,1).It is important to keep in mind that
the notation (0,1) is used to signify the second sample along the first row. It does not mean that these are
the actual values of physical coordinates when the image was sampled. Following figure shows the
coordinate convention. Note that x ranges from 0 to M-1 and y from 0 to N-1 in integer increments.

The coordinate convention used in the toolbox to denote arrays is different from the preceding
paragraph in two minor ways. First, instead of using (xylem) the toolbox uses the notation (race) to
indicate rows and columns. Note, however, that the order of coordinates is the same as the order discussed
in the previous paragraph, in the sense that the first element of a coordinate topples, (alb), refers to a row
and the second to a column. The other difference is that the origin of the coordinate system is at (r, c) = (1,
1); thus, r ranges from 1 to M and c from 1 to N in integer increments. IPT documentation refers to the
coordinates. Less frequently the toolbox also employs another coordinate convention called spatial
coordinates which uses x to refer to columns and y to refers to rows. This is the opposite of our use of
variables x and y.

Data Classes:

Although we work with integers coordinates the values of pixels themselves are not restricted to be
integers in MATLAB. Table above list various data classes supported by MATLAB and IPT are
representing pixels values. The first eight entries in the table are refers to as numeric data classes. The
ninth entry is the char class and, as shown, the last entry is referred to as logical data class.

All numeric computations in MATLAB are done in double quantities, so this is also a frequent data
class encounter in image processing applications. Class unit 8 also is encountered frequently, especially
when reading data from storages devices, as 8 bit images are most common representations found in
practice. These two data classes, classes logical, and, to a lesser degree, class unit 16 constitute the
primary data classes on which we focus. Many ipt functions however support all the data classes listed in
table. Data class double requires 8 bytes to represent a number uint8 and int 8 require one byte each,
uint16 and int16 requires 2bytes and unit 32.

Image Types:

The toolbox supports four types of images:

1 .Intensity images

2. Binary images

3. Indexed images

4. R G B images

Most monochrome image processing operations are carried out using binary or intensity images, so
our initial focus is on these two image types. Indexed and RGB colour images

Intensity Images:

An intensity image is a data matrix whose values have been scaled to represent intentions. When the
elements of an intensity image are of class unit8, or class unit 16, they have integer values in the range
[0,255] and [0, 65535], respectively. If the image is of class double, the values are floating _point
numbers. Values of scaled, double intensity images are in the range [0, 1] by convention.
Binary Images:

Binary images have a very specific meaning in MATLAB.A binary image is a logical array 0s
and1s.Thus, an array of 0s and 1s whose values are of data class, say unit8, is not considered as a binary
image in MATLAB .A numeric array is converted to binary using function logical. Thus, if A is a numeric
array consisting of 0s and 1s, we create an array B using the statement.

B=logical (A)

If A contains elements other than 0s and 1s.Use of the logical function converts all nonzero
quantities to logical 1s and all entries with value 0 to logical 0s.

Using relational and logical operators also creates logical arrays.

To test if an array is logical we use the I logical function:

islogical(c)

If c is a logical array, this function returns a 1.Otherwise returns a 0. Logical array can be converted
to numeric arrays using the data class conversion functions.

Indexed Images:

An indexed image has two components:

A data matrix integer, x.

A color map matrix, map.

Matrix map is an m*3 arrays of class double containing floating_ point values in the range [0, 1].The
length m of the map are equal to the number of colors it defines. Each row of map specifies the red, green
and blue components of a single color. An indexed images uses “direct mapping” of pixel intensity values
color map values. The color of each pixel is determined by using the corresponding value the integer
matrix x as a pointer in to map. If x is of class double ,then all of its components with values less than or
equal to 1 point to the first row in map, all components with value 2 point to the second row and so on. If
x is of class units or unit 16, then all components value 0 point to the first row in map, all components
with value 1 point to the second and so on.

RGB Image:

An RGB color image is an M*N*3 array of color pixels where each color pixel is triplet
corresponding to the red, green and blue components of an RGB image, at a specific spatial location. An
RGB image may be viewed as “stack” of three gray scale images that when fed in to the red, green and
blue inputs of a color monitor
 Produce a color image on the screen. Convention the three images forming an RGB color image are
referred to as the red, green and blue components images. The data class of the components images
determines their range of values. If an RGB image is of class double the range of values is [0, 1].

Similarly the range of values is [0,255] or [0, 65535].For RGB images of class units or unit 16
respectively. The number of bits use to represents the pixel values of the component images determines
the bit depth of an RGB image. For example, if each component image is an 8bit image, the corresponding
RGB image is said to be 24 bits deep.

Figure 3.4 : An image is an array or a matrix of pixels arranged in columns and rows.

In a (8-bit) grayscale image each picture element has an assigned intensity that ranges from 0 to 255. A
grey scale image is what people normally call a black and white image, but the name emphasizes that such
an image will also include many shades of grey.

Pixel :

A pixel (short for picture element) is a single point in a graphic image. Each such information element is
not really a dot, nor a square but an abstract sample. Each element of the above matrix is known as a pixel where
dark = 0 and light = 1. A pixel with only 1 bit will represent a black and white image. If the numbers of bits are
increased then the number of gray levels will increase and a better picture quality is achieved. All
naturally occurring images are analog in nature. If the number of pixels is more than the clarity is more.
An image is represented as a matrix in Digital Image Processing. In Digital Signal Processing we use only
row matrices. Naturally occurring images should be sampled and quantized to get a digital image. A good
image should have 1024*1024 pixels which is known as 1k * 1k = 1Mpixel
Figure 3.5: Each pixel has a value from 0 (black) to 255 (white). The possible range of the pixel values
depend on the color depth of the image, here 8 bit = 256 tones or grayscales.

A normal grey scale image has 8 bit color depth = 256 grey scales. A “true color” image has 24 bit color
depth = 8 x 8 x 8 bits = 256 x 256 x 256 colors = ~16 million colors.

Figure 3.6: A true-color image assembled from three grayscale images colored red, green and blue. Such
an image may contain up to 16 million different colors. Some grayscale images have more grayscales, for
instance 16 bit = 65536 grayscales. In principle three grayscale images can be combined to form an image
with 281,474,976,710,656 grayscales.

1.3 RGB image

The RGB color model relates very closely to the way we perceive color with the r, g and b receptors in
our retinas. RGB uses additive color mixing and is the basic color model used in television or any other
medium that projects color with light. It is the basic color model used in computers and for web graphics,
but it cannot be used for print production.

The secondary colours of RGB – cyan, magenta, and yellow – are formed by mixing two of the primary
colours (red, green or blue) and excluding the third color. Red and green combine to make yellow, green
and blue to make cyan, and blue and red form magenta. The combination of red, green, and blue in full
intensity makes white. The RGB color model is an additive color model in which red, green,
and blue light are added together in various ways to reproduce a broad array of colors. The name of the
model comes from the initials of the three additive primary colors, red, green, and blue.

The main purpose of the RGB color model is for the sensing, representation, and display of images in
electronic systems, such as televisions and computers, though it has also been used in
conventional photography. Before the electronic age, the RGB color model already had a solid theory
behind it, based in human perception of colours.

RGB is a device-dependent color model: different devices detect or reproduce a given RGB value
differently, since the color elements (such as phosphors or dyes) and their response to the individual R, G,
and B levels vary from manufacturer to manufacturer, or even in the same device over time. Thus an RGB
value does not define the same color across devices without some kind of color management. Typical
RGB input devices are color TV and video cameras, image scanners, and digital cameras. Typical RGB
output devices are TV sets of various technologies (CRT, LCD, plasma, etc.), computer and mobile
phone displays, video projectors, multicolor LED displays, and large screens such as JumboTron, etc.
Color printers, on the other hand, are not RGB devices, but subtractive color devices (typically CMYK
color model).

Figure 3.7: The additive model of RGB. Red, green, and blue are the primary stimuli for human color
perception and are the primary additive colors

1.3.1 Astronomical images

Images of astronomical objects are usually taken with electronic detectors such as a CCD (Charge
Coupled Device). Similar detectors are found in normal digital cameras. Telescope images are nearly
always gray scale, but nevertheless contain some color information. An astronomical image may be taken
through a color filter. Different detectors and telescopes also usually have different sensitivities to
different colors (wavelengths).

1.3.2 Representative color images

If one or more of the images in a data set is taken through a filter that allows radiation that lies outside the
human vision span to pass – i.e. it records radiation invisible to us - it is of course not possible to make a
natural colour image. But it is still possible to make a colour image that shows important information
about the object. This type of image is called a representative color image. Normally one would assign
colors to these exposures in chromatic order with blue assigned to the shortest wavelength, and red to the
longest. In this way it is possible to make color images from electromagnetic radiation far from the human
vision area, for example x-rays. Most often it is either infrared or ultraviolet radiation that is used.
Fundamental steps in Digital Image Processing

Image Acquisition:

Digital image acquisition is the creation of digital images typically from a physical object. A digital image
may be created directly from a physical scene by a camera or similar device. Alternatively it can be
obtained from another image in an analog medium such as photographs, photographic film, or printed
paper by a scanner or similar device. Many technical images acquired with tomographic equipment, side-
looking radar, or radio telescopes are actually obtained by complex processing of non-image data.

Image Enhancement:

The process of image acquisition frequently leads to image degradation due to mechanical problems, out-
of-focus blur, motion, inappropriate illumination and noise. The goal of image enhancement is to start
from a recorded image and to produce the most visually pleasing image.

Image Restoration:

The goal of image restoration is to start from a recorded image and to produce the most visually pleasing image.
The goal of enhancement is beauty. The goal of restoration is truth. The measure of success in restoration is
usually an error measure between the original and the estimate image. No mathematical error function is
known that corresponds to human perceptual assessment of error

Color image processing:

Color image processing is based on that any color can be obtained by mixing 3 basic colors red, green and
blue. Hence 3 matrices are necessary each one representing each color.

Wavelet and Multi-Resolution Processing:

Many times a particular spectral component occurring at any instant can be of particular interest. In these
cases it may be very beneficial to know the time intervals these particular spectral components occur. For
example, in EEGs the latency of an event-related potential is of particular interest. Wavelet transform is
capable of providing the time and frequency information simultaneously, hence giving a time-frequency
representation of the signal. Although the time and frequency resolution problems are results of a physical
phenomenon (the Heisenberg uncertainty principle ) and exist regardless of the transform used, it is
possible to any signal by using an alternative approach called the multi-resolution analysis (MRA). MRA
analyzes the signal at different frequencies with different resolutions. MRA is designed to give good time
resolution and poor frequency resolution at high frequencies and good frequency resolution and poor time
resolution at low frequencies.

Compression:
Image compression is the application of data compression on digital images. Its objective is to reduce
redundancy of the image data in order to be able to store or transmit data in an efficient form.

Morphological processing:

Morphological processing is a collection of techniques for image processing based on mathematical

morphology. Since these techniques rely only on the relative ordering of pixel values not on their
numerical values they are especially suited to the processing of binary images and gray scale images.

Segmentation:

In the analysis of the objects in images it is essential that we can distinguish between the objects of
interest and ³the rest´. This latter group is also referred to as the background. The techniques that are used
to find the objects of interest are usually referred to as segmentation techniques.

In image processing it is usually necessary to perform high degree of noise reduction in an image before
performing higher-level processing steps, such as edge detection. The median filter is a non-linear digital
filtering technique, often used to remove noise from images or other signals. The idea is to examine a
sample of the input and decide if it is representative of the signal. This is performed using a window
consisting of an odd number of samples. The values in the window are sorted into numerical order; the
median value, the sample in the center of the window, is selected as the output. The oldest sample is
discarded, a new sample acquired, and the calculation repeats.

Median filtering is a common step in image processing. It is particularly useful to reduce speckle noise
and salt and pepper noise. Its edge-preserving nature makes it useful in cases where edge blurring is
undesirable. Image synthesis is the process of creating new images from some form of image description.
The kinds of images that are typically synthesized include:

test Patterns - scenes with simple two dimensional geometric shapes.

Image Noise - images containing random pixel values, usually generated from specific parameterized
distributions.

Computer Graphics - scenes or images based on geometric shape descriptions. Often the models are three
dimensional, but may also be two dimensional.

Synthetic images are often used to verify the correctness of operators by applying them to known images.
They are also often used for teaching purposes, as the operator output on such images is generally `clean',
whereas noise and uncontrollable pixel distributions in real images make it harder to demonstrate
unambiguous results. The images could be binary, grey level or colour.

1.4Tools And Techniques Of Image Processing

Numerous numbers of sources will be a reason for the addition of unwanted signs in binary images in the
form of the unhealthy illumination and noise [86-94]. This happens especially when we are acquiring,
quantization and transmitting the images.

The concept of noise is unavoidable in image progressioning. The degradation will directly affect the
visualization of the binary images when we are progressioning the image. This is still an unsolved
problem which has a wide capability of study in this area takes us to introduce a scheme to answer this
issue.

At present modernization of images will have an imperative part in every aspect especially in the
exchange of data which having certain details in it. Images and videos are used all over in the world to a
greater extent depending on their usage. The unwanted sign will make change in the true image at edges,
lines; get ridding of the image details and pixel values. This is the main reason for removing of noise in
the images before progressioning the image. The different progressioning techniques will be discussed in
the further section in below.

1.5 Noise removal of image

In present and future scenario will be very much related to the digital image progressioning which having
the integral part of every one’s life. The applications which always turn around the extended general
documentations of a probability and visual exchange of information medical fields and serious
surveillance, the quality of the image and accuracy is increased along with demand that have for images.

The blur in the images will be attained while acquiring the image and transmitting of pictures which are
degraded quality, which spoiled the picture quality in both sensing and visual awareness. The quality also
depends on the environmental conditions when the image has been captured. For example, the images that
have been sending by the wireless medium maybe be affected by the lightening effects, temperature
variation etc.

Get rid of the image is a well explored topic that have been prime aim to our any progressioning the image
which enhances quality of image by diminishing the noise from the image. The main theme here to
safeguard the required information of pictures and removing undesired details on the image, which has to
be achieved by the local geometrics and pixels values having imperative data [92- 94].

Denoised
Input image Noise image Noise
image
f(x, y) + g(x, y) reduction 
window f (x, y)

Unwanted
signal
n (x, y)
Fig.3.8 A image model removing of noise progression

From the above shown figure 3.8 where the input image subjected to the addition of noise n (x, y) and
noise image g(x, y) provides the required output. Which has been given to the noise reduction filter (s)

and the denoised image is obtained at the output ( f (x, y) ).

1.6 Image Noise

The noise is introduced when we are considering an image, formation, recording, or transmitting section,
which has to be trim down completely using various techniques. Even a minute amount of unwanted sign
will change the whole scenario, when accuracy is needed. The unwanted signs are mainly of following
types. They are impulse, additive white Gaussian unwanted signal (AWG), unwanted speckle signal,
Poisson, salt and pepper blurs and many more.

A vast number of techniques have been future by number of authors which paved the way for the further
development in the removing of noise concept. These techniques will further be categorized on their
nature and use.

Here various types of noises are discussed that have encountered in the proposed technique. They are
shown below.

1.6.1 Noises types and sources

Material dimension is one dependent for the source and types of noise. Which is likely to be dependent on
then quantity that to be measured. The unwanted sign is attained when the source is different from the
required source; even when the noise is caused form the measuring progression itself.

Sometimes the unwanted sign occurs due to mathematical iteration also, especially when we try to
compress an image. Noise is undesirable part of all the digital images. It is not possible to delete the noise
that has been in the digital images when we are acquiring the image so it is essential to delete the
unwanted signal by using of noise algorithms.

1.6.2 Noise in Camera

From variety of sources we will see the noise when we are taking the images from the digital cameras and
conventional film cameras. These noises are mainly huge in number. They were listed here as follows.

The Gaussian, salt and pepper blur were two frequently widespread unwanted signals that we observe. In
every case the unwanted signals various pixels are correlated or uncorrelated. These may be independent
or the identically separated. The most frequently occurred noises are seen here they are:
1.6.3 Additive White Gaussian blur (AWG)

The most commonly used noise that has been seen in the black and white images which are frequently
used in the digital image progressioning. These are mainly used as a benchmark standard where the
AWGN will assess the performance of the image using removing of noise algorithms. The probability
distribution function of normal distribution (Gaussian distribution) is due to Gaussian noise, this noise is
also called as statistical noise.

Normally dram rate is extra to the spotless pel value in Gaussian distribution, this will clearly be observed
in Gaussian phenomenon. This is exactly same for all the pixels (mean = variance). Here the amount of
noise is added to each and every value of the pixel where the amplifier noise of every noise is same as the
AWGN. This Gaussian noise is mainly occurred due to the thermal agitation of the charged carriers which
is often seen within the electrical conductor. The bell-shaped function will show the distribution function
of the Gaussian variable (bell curve) is exposed in fig 3.9.

Fig.3.9 The distribution of Gaussian function (bell curve)

In general statistics the normal distribution is taken into account most prominently. The main reasons for
this are the central limit theorem comes under the mild conditions, the huge number of random variables
is considered independently which is taken into account from same distribution. The theorem is tactical
analytically which is derived explicitly.

1 f   2
1 ( )
Pf(f) e2 
 2

Where  is average/expectation (site of the crest value?)

  2 is variance

 is standard deviation

1.6.4 Salt and Pepper blur

It has distribution of fat tail, the other name is spike blur or the impulsive blur. The main reason for this
noise is due to the loss of the pixel while we transmit the images. Here when we compared the pixels with
the surrounding images then we can notice the variation in those pixels in color or intensity of those
pixels, but is very limited to a small number of pixels. Here the noise type will make the image either dark
or saturated. This makes the rate of the pixels is uncorrelated with rate of the pixel in spotless picture. This
occurs mainly due to the dust in the camera or overheated faulty CCD elements, A-D convertor errors etc.
The bit errors in transmission are also caused due to this. The removing of noise effect of median filters is
quite being effective on this kind.

1.6.5 Photon Shot Noise

This is mainly occurred by quantum fluctuation which is dominant in lighter parts of an image for which
we have taken it from the image sensor; this describes he change in the photon count which sensed at a
given exposure level. This having the root average square number that is directly relative to the intensity
square root and independent to another. This also backups the poisons noise with its distribution, which do
not vary far away from the Gaussian too. If the mean is too large then it is to be analogous like the
Gaussian distribution having Mean and Variance. The deleting of the Gaussian and the Poisson noise is
the same. This has become the major concern in the recent times due to the image sensor is reducing the
size and the also decrease in the number of the photons that is capturing the image.

1.6.6 Thermal noise

This caused mainly due to the emery of the chip that will accumulate the photoelectrons resulting
in the unwanted signs. This also occurs in the absence of the light and hence called as the dark-current
noise which varies accordingly the temperate of the sensor. This can diminish by the sign acquiring
progression which can be done by the cooling the sensor of the camera.

1.7 Other Types of Noises

1.7.1 Noises that are spatially independent

Here all the spatial coordinates of the blur is independent as well as uncorrelated according to the picture
itself, applications like X-rays and nuclear image/imaging on medicine. The 1.1 table demonstrates the
spatial self-governing PDF values the variances and mean.

1.7.2 Blurs those are spatially dependent

 Here the type of blur will be depends on the spatially correlated itself and coordinates to figure
itself. Here the most frequently noise that occurs is periodic noise which is correlated to it. This periodic
noise occurs mainly due to the electrical and the electromagnetic interferences which acquiring the image.

Table 3.1 Probability Density Functions PDF’s blur categories, Average and Variance

Blur type PDF Average Variance

 (s  a)2
1
p(z)  e 2b2
,   z  
2 b ma
Gaussian  2  b2

 (s  a)2
2
p(z)   (z  a) e b
,z a
b
Rayleigh b b(4   )
ma 2 
za 4 4
=0,

 a b z b 1  az
p(z)   e
 (b  1)! ,z0
Erlang b b
m 2 
a a2
=0 , z0
(gamma)

p(z)   ae az , z  0
Exponential 1 1
=0, z0 m 2 
a a2

1
p (z) 
ba ,a  z  b
Uniform ab (b  a) 2
m  
2
=0, elsewhere 2 12

Salt and p a , z  a

pepper p (z)   pb , z  b
0, otherwise m  a p a  bpb  2  (a  m)2 p a  (b m)2 pb

(impulse)

1.8 Demonstration of Image

The human visual system (HVS) incurs in given picture like an accumulation of Luminant energy which
is distributed spatially [20]. This type of representation termed as optical image. Optical images are
general types of images that a camera captures. Later these are constituted as video like an analog
electrical signal. These are further sampled to transform in to a digital image which is represented as S(r ,
c).

The binary image S(r, c) is interpreted like 2D collection of information. This is called as Pixel
representation. Each pixel magnitude refers to certain feature of the image in spatial system. One such is
the brightness of the image. Finally, the whole image is transformed into a data matrix having rows and
columns to fit 2D array representation. Every row in the matrix matches for a vector, based on mapping
many image formats [95-104].

1.8.1 Binary Image

The elementary type of image representation is called the "Binary image". It typical uses only two levels.
The two levels are referred to as black and white which are mentioned as ‘1’ and ‘0’. This kind of image
representation is considered like 1 bit per pixel image. This is suitable to an reason which considers barely
single digit number to signify every pel. These types of images are often used to depict low level
information of the picture like its outline or shape. Especially in applications like representation in optical
character (OCR) where the only outline character required realizing the letter representing it. The digital
images are generated from the system of images of gray scale through a technique called thresholding.
The two-level thresholding simply acts as a decision factor above which it switches to numerical '1' and
below which it switches to numerical '0'.

(a) (b)
(c)

Fig. 3.10 illustrations of digital images

1.8.2 Image of Gray scale

Images of Gray scale (GS) were denoted as neutral or single-color picture. Such images possess
information of brightness merely. Hence color data is contained by them is empty. However, the
brightness is represented at different levels. Typical 8-bit image holds a range of 0-255 brightness levels
known as gray levels. Here 0 refers to black and 1 refers to white. The 8-bit depiction is obvious with the
reality to a computer actually handles the data in 8-bit format. Below Fig.3.11 (a) and (b) are two
examples of such GSI.

(a) (b)
(c)

Fig.3.11 Examples of GSI

1.8.3 Color images

Color image (CI) which modeled as triple band single chromatic light information, here every band of
information will keeps in touch with various dissimilar colors.

(a) (b)

(c)
Fig.3.12 illustrations of CI

The genuine data accumulated via binary data of the image is brightness in all spectrum bands. The
displaying of the image that corresponds sequence of brightness which depicts on the screen by image
components that gives out energy in the form of light that corresponds to a singular color. Colored
pictures that are distinctive were signified as RGB images that uses the eight bit monochrome ordinary as
a replica, the color images that correspond will have 24 bits to 8 bits per pixel for each color bands (red,
green, blue).

(a) SR(r,c) (b) SB(r,c) (c) SG(r,c)

Fig.3.13: Representation of CI as individual color (RGB) images.

The following figure illustrates the vector as an arrow that adds, which refers an individual smallest unit
measurement of red, green, blue principles like a color vector ---(R, G, B ).

Fig.3.14 A shading pixel vector comprises of the red, green and blue pixel esteems (R, G, B) at one given
line/section pixel facilitate (r, c)
Fig.3.15 Color pixel representation

For many applications, the information about RGB shading is changed into mathematical space that
decouples the brilliance information from the genuine color data. The HSL shading change enables we to
portray shades by means of all promptly get it. The daintiness is the splendor of the shading, and the tint is
the thing that we regularly consider as "shading" and the tone (ex: green, blue, red, and orange). The
immersion is a proportion of how much white is in shading (ex: Pink equals to red when it has extra white,
so it is less soaked than an unadulterated red).

1.8.4 Multispectral Images

A multispectral picture is one that catches picture information at particular frequencies over the
electromagnetic range. Multispectral pictures regularly contain data outside the typical human perceptual
range. This may incorporate infrared, bright, X-beam, acoustic or radar information. Foundation of these
kinds of picture incorporates satellite frameworks, submerged sonar frameworks and medicinal
diagnostics imaging frameworks.
Fig3.16 Electromagnetic range

(a) (b)

Fig.3.17 Examples of Multispectral images

1.8.5 Computer Graphics

PC illustrations are a specific field inside that alludes to the software engineering domain that alludes to
the multiplication of visual information using PC.

In PC illustrations, sorts of picture information are separated into two principle classes:

1. Bitmap picture (or raster picture): can spoke to by our picture demonstrate I(r, c), where we have pixel
information and comparing brightness esteems that put away in file arrangement.

2. Vector pictures: allude to the strategies for shows lines, bends shapes by storing only just the key
focuses. These key focuses are adequate to characterize the shapes, and the procedure of turning
postulations into a picture is called ripping after the picture has been rendered, it very well may be thought
of as being in bit outline where every pixel has particular qualities related with it.

1.9 Digital Image Formats

Picture groups are institutionalized methods for sorting out and putting away advanced pictures [101].
Picture records are taken from advanced data in individual configurations that can be rasterized for use on
a PC show or printer. A picture manuscript configuration may stockpile data in uncompressed, full to
capacity, or designs of vectors. Once rasterized, an image turns into a framework of pixels, every one of
which has different bits to allocate its shading equivalent to the coloring profundity of the gadgets shows
it. Most the sort of document organize fall into class of bitmap pictures. By and large, these sorts of
pictures contain both header data and the crude pixel information. The header data contain data with
respect to

i. The numeral rows (height)

ii. The numeral columns (Width)

iii. The numeral groups.

iv. The quantity of bit per pixel.

v. The document type

vi. Moreover, with a portion of the beneficial complex record designs have the header may contain
information about the kind of pressure utilized and other vital requirements to make the picture, I (r, c).

BMP arrange (Bitmap picture File Format)

The BMP otherwise called bitmap picture record or gadget free bitmap (DIB) document arrange or
basically a bitmap, is a raster illustrations picture document design used to stock up bitmap computerized
pictures, independently of the showcase gadget, (for example, a designs connector), specifically on
Microsoft Windows and OS working frameworks.

This record arrange can store 2D double pictures of arbitrary width, tallness, and goals in both
monochrome and shading pictures, in an assortment of shading profundities, and alternatively with
information demands, alpha channels, and shading plots.

TIFF (Tagged Image File Format)

Labeled Image File Format is a standout amongst the most famous and adaptable of the present open
space raster document positions. They are utilized on World Wide Web (WWW). GIF records are
restricted to a most extreme of 8 bits/pixel and takes into consideration a kind of pressure called LZW.
The GIF picture header is 13 byte long and contains essential data.

JPEG (Joint Photo Graphic Experts Group)

This is the correct organization for those photograph pictures which must be little records, for instance, for
sites or for email. JPG is frequently utilized on advanced camera memory cards. The JPG record is
magnificently little, frequently packed to maybe just 1/10 of the span of the first information, which is
something to be thankful for when modems are included. In any case, this awesome pressure effectiveness
accompanies a high cost.

JPG utilizes lossy pressure (lossy signifying "with misfortunes to quality"). Here Lossy implies that nature
of the picture is lost where the pressure of JPG information is done and spared, and the recuperation of
this quality is never to be occur. JPEG pictures pressure is being utilized widely on the WWW. It's,
adaptable, so it can make huge documents with fantastic picture correspondence.

VIP (visualization in picture handling) positions:

It is produced for the CVIP devices programming, when performing brief pictures are made that
utilization skimming point portrayal which is past the standard 8 bit/pixel. To speak to this sort of
information the remapping is utilized, which is the way toward taking unique picture and adding a
condition to make an interpretation of it to the rang (0-225).

Graphics Interchange Format (GIF)

The indexed representation of color images is used by GIF (having the palette of maximum 256 shades),
the LZW (Lempel–Ziv–Welch) firmness procedure, and title of size 13 byte.

Portable Network Graphics (PNG)

The most popular image type of format which supports both true color and indexed images is PNG.
Additionally, it offers a patent-free alternate for GIF set-up.
CHAPTER 4

MATERIAL AND METHODS

4.1 Soil sample collection

The proposed soil classification system contains 50 soil samples. Samples are collected from 10
paddy fields of west Guwahati under the supervision of the Department of Civil Engineering,
Girijananda Chowdhury Institute of Management and Technology (GIMT), Guwahati, Assam,
India. The sample collection site lies in between 26.1445° N latitude and 91.7362° E longitude.
Five samples are collected from each field with a separation of 200 m. Soils are dragged in a
depth of 6 in. from the top-level of the paddy field and kept in a white paper for image
acquisition. Hydrometer test is used to calculate the fraction of sand, silt, and clay in the soil
and determined the texture of the soil using USDA triangle. In Fig. 4.1, the area under the black
circle represents the site of sample collection.

4.2 Image acquisition

Image acquisition is the process of capturing soil images from the samples. A Xiaomi Redmi 3S
smartphone with the marsh- mallow version of Android is used to photograph the soil images.
The Smartphone camera consists of a 13-megapixel charged-coupled device. During the time of
image capture, all the camera properties are kept in default settings such as F-stop = f/2,
exposure time = 1/60 S, ISO speed = ISO 125,
Fig. 4.1 – Area for Sample Collection

focal length = 4 mm. Fig. 4.2 shows the procedure of the image acquisition. Soil samples are
photographed in a closed room with only one curtain less window. Light enters into the room
through the window. The effect of natural light is less in this process and is also very cost
effective. In order to avoid the other environmental light condition, soil images are pho-
tographed from 10 in. ahead from the soil in a straight position as shown in Fig.4.2. All the soil
images are kept as per field starting with a number from 1 to 5. Five samples are pho- to graphed
from each field and a total of 50 soil samples are found using an Android Smartphone. Fig. 4.3
shows the samples of the captured soil images

4.3 Soil texture determination in laboratory

In this section, the procedure of soil texture classification using the hydrometer and USDA
triangle methods are explained. Soil consists of sand, silt, and clay [10]. The fraction of sand,
silt, and clay gives an idea of the soil texture and are calculated using the pipette or hydrometer
method [10]. The final classification is done using USDA criteria. The proposed system chose
the hydrometer method as the labora- tory method of soil texture classification as it clearly gives
the fraction of sand, silt, and clay present in a soil sample [11]. The other reason for choosing
the hydrometer method is that

Fig. 4.2 Soil Sample Collection and its ROI after Segmentation
Fig 4.3 Some of the samples of Collected Soil Image Database.

it is available in GIMT, Guwahati, and nearest to the sample collection site.

The hydrometer test is done in the civil engineering department of GIMT, Guwahati, Assam
India. The hydrometer is calibrated by preparing a solution of 28-gram tablespoon sugar and 176
g of distilled water. The solution is prepared up to the calibration temperature of the hydrometer
and measure the gravity of the solution to 1.048. After the calibration, all the soil samples are
processed for hydrometer test. In the lab, 50 g of each soil sample within a limit of 2 mm coarse
fragment are dispersed with sodium metaphosphate and then agitated. The entire process follows
all the steps of the hydrometer test [11]. At last, the fraction of sand, clay, and silt is calculated
using the following equations [11].

% Clay ¼ Corrected hydrometer reading at 6 hrs; 52 min

× 100=wt of sample

% Silt ¼ Corrected hydrometer readings at 40 sec

× 100=wt of sample — % of clay

% Sand ¼ 100% — %Silt — %Clay

The result of the hydrometer method is recorded in the percentage of sand, silt, and clay. The
textures of the soil sample are determined using the United State Department of Agriculture
(USDA) triangle method. For the implementation, a Mat lab code is developed to classify the
texture of the soil and the output of the USDA classification is represented in the excel sheet.
The actual soil texture of sample 1 from the first paddy field is shown in Fig. 4.4 as per the
USDA criteria.

In Fig. 4.4, the red point is located in the area of silt loam. Loam soil means the combination of
sand, silt, and clay. Silt loam expresses all the fraction of three end members of the

Fig. 4.4 USDA Triangle Value for Sample 1

soil, but the fraction of silt is more as compared to the other two one.

In the hydrometer test for 50 soil samples, it is found that

A total of 4 samples out of 50 samples are clay and these belong to paddy fields 3 and 9.

A total of 7 samples out of 50 samples are clay loam and these belong to paddy fields 3, 5, 8 and
9.

Only 2 samples out of 50 samples are fine sand and these belong to paddy field 2.

Only 2 samples out of 50 samples are loam and these belong to paddy field 5

Table 1 – Mean Fraction of Sand, Silt, and Clay of each soil Sample.
Mean % Sand Mean % Silt Mean % Clay Texture (USDA)
23.51 15.38 61.12 Clay
31.87 36.39 31.72 Clay Loam
95.00 0.00 2.50 Find Sand
38.00 38.13 23.81 Loam
85.00 5.00 10.00 Loamy Fine Sand
85.00 3.00 12.00 Loamy Sand
54.64 6.40 39.08 Sandy Clay
61.52 12.58 26.19 Sandy clay loam
19.15 63.68 17.16 Silt Loam
20.00 40.00 40.00 Silty Clay
13.21 53.95 32.83 Silty Clay Loam

Fig. 4.5 – Fraction of Clay Soil Sample.

Fig. 4.6 – Fraction of Fine Sand Soil Sample.

Fig. 4.7 – Fraction of Clay Loam Soil Sample

Fig. 4.8 – Fraction of Loam Soil Sample.

Only 1 sample out of 50 samples is loamy fine sand and it belongs to paddy field 4.

Only 1 sample out of 50 samples is loamy sand and it belongs to paddy field 6.

A total of 3 samples out of 50 samples are sandy clay and these belong to paddy fields 4 and 9.

A total of 10 samples out of 50 samples are sandy clay loam and these belong to paddy fields 2,
4, 6, 7 and 9.

After the hydrometer method, USDA classification gives 12 different types of soil classes
including Clay, Clay Loam, Loam, Fine Sand, Sandy Clay loam, etc. The samples of each class
are divided into training and testing set. It is presented

Table 2 – Description of training and testing

samples of soil image.
Soil class Total Training Testing
 Number ofImage Images
Samples
Clay Loam 7 4 3
Sandy Clay10 5 5
Loam
Silt Loam 11 6 5
Clay 4 2 2
Fine Sand 2 1 1
Loam 2 1 1
Loam Fine1 1 1
sand
Loamy Sand 1 1 1
Sandy Clay 3 2 1
Sandy Loam 5 3 2
Silt Clay 1 1 1
Silt Clay Loam 3 2 1
Total 50 29 24
in Table 2. Among all the soil classes, 3 soil classes such as Fine sand, Loamy sand, and Silt
Clay soil have only 1 image. So, the same image is considered for training and testing.

4.4 Soil image Pre-processing, segmentation and feature extraction

4.4.1 Pre-processing

Pre-processing of soil image means the enhancement of soil image quality for better feature
extraction [38] and texture classification. Since the intention is to evaluate the soil
classification using an Android Smartphone, soil images are collected using a Smartphone.
The dimensions of the images are in 4160 × 2630 pixels with an average size of 2.3 MB.
The contrasts of the images are enhanced using Mat Lab 2017a. The output of the contrast-
enhanced image is presented in fig 4.9. Then the images are processed for noise removal and
the low pass butter worth filter with a cut off frequency 0–0.5 is used to remove the noise of
¼
the images.
Fig. 4.9 – Image Pre-processing and Segmentation

Above Eq. (iv) shows the butter worth low pass filter for noise remove where n is the order of
filtering. The order of the filtering is 2 in our approach. After the noise removal, the images are
resized into a dimension of 300x400 and are processed for the extraction of the region of
interest. The algorithm for image pre-processing is presented by Algorithm 1.

Algorithm 1: Algorithm for the Image Pre-

processing
Input: Input Image

Output: Pre-processed Image

Step1: Read the Images

Step 2: Enhanced the Contrast of the Image.

Step 3: Remove the noise of the Contrasted

Image using Low pass

Butter worth Filter.

Step 4: Images are resized in a dimension of

300x400.

IMAGE SEGMENTATION
All the filtered images are with white background region. Though the images are with white
background, the region of interest is extracted from the soil image using segmentation. Image
segmentation is the process of segment an image into the different region and selects the region
of interest of an image from it. It is found that the Gabor filter is a good model for texture
segmentation of an image to identify the region of interest. For this approach, an array of Gabor
filter is used with a different scale and orientations [12]. The scale and orientation of the Gabor
filter for this approach are 4 and 6 respectively. For the implementation, a Mat lab (R2017a)
code is developed and the output of the code is figured out in Fig. 16. The first part of Fig 4.9
represents the captured image, the second part of the Fig. 16 represents the contrast-
enhanced image and the last part of the Fig. 4.9 rep- resents the segmented image.

FEATURE EXTRACTION OF IMAGES

Feature extraction is the process of extraction of different texture features of a soiled image. The
extracted features of the image are presented as a feature vector of the soil images. The following
steps are followed for the feature extraction and the method is presented in Fig. 17.

In the first step, the Hue, Saturation, and Value (HSV) histogram [31] of the segmented images
are calculated. The input image is quantized into Hue, Saturation, and Value (HSV) color space
with 8 × 2 × 2 equal bins. The output of histogram is a vector of size 1 × 32 these con- tains the
Hue, Saturation, and Value of the images.

In the second step, the color auto correlogram is calculated. It is used to find the spatial
correlation of the identical pixels [23,31,32]. To implement color auto correlogram using Mat
Lab 2015a, the input image is quantized into 64 colors in RGB color space as 4 × 4 × 4 color
space. The output is presented as a feature vector of size 1 × 64 containing color auto
correlogram. This method follows a dynamic programming approach with a distance set d = {1 3
5 7} as suggested by Hung et al. [23].

In this step, the color moment of the input image is determined. It is used to find the color
similarity between the images [24,31,32]. In this paper, the RGB color channel of the input
image is considered and computed the 9 color Moments of the original image out of these 3 are
for mean, 3 are for standard deviation [31] and 3 are for skewness. These moments are for the
Fig.4.10 – Flow Chart of Feature Extraction

three channels of the images [13,29,32]. The output of this method is represented as a feature
vector of size 1x9.

In the fourth step, the Gabor feature of the soil images is calculated. Gabor wavelet is a method
to find the Gabor feature of an input image [12,31,32]. This method is used to calculate the
mean-squared energy and mean amplitude of the input image. For this paper, the number of
wavelet scale is 4 and the number of filter orientation is 6. Before applying this method, images
are converted into a grayscale format and perform the convolution using the First Fourier
transformation. The output size of the mean-squared and mean-amplitude feature of the Gabor
filter is 1 × 24 and 1 × 24 respectively. The final size of the output feature of the Gabor filter is 1
× 48.

In the next step, the single level discrete 2-D wavelet decomposition [32] is performed. The
wavelets coefficients of the input image are extracted. The feature vector contains the mean and
standard coefficients of the discrete wavelet transform. The output size of the mean and standard
coefficient feature of the discrete wavelet transform is 1 × 20 and 1 × 20 respectively. The final
size of the output feature of the Discrete Wavelet Transform is 1 × 40.

In the last step, all the calculated features are combined together to form the final feature vector
of the image. The output of the combination is presented as the final feature vector of the
methodology.
Initially, the multi SVM classifier is used to classify 3 different soil classes and later on, the
multi SVM classifier is used to classify 12 different soil classes. The feature vector size of one
single image is 1 × 193. The final feature vector size for 15 training images of 3 class soil
classifier is 15 × 193.
The size of the testing set is 13 × 193 for the entire 13 images of 3 class classifier. The final
feature vector size for the entire 29 training images of 12 class classifier is 29 × 193 and the size
of the testing set is 24 × 193.

CLASSIFICATION, RESULT DISCUSSION AND ANALYSIS

Table 1 shows the percentage of sand, silt, and clay of different soil samples and results are
displayed in terms of texture using USDA triangle method. It means that the hydrometer test
gives only the percentage of sand, silt, and clay in the lab- oratory, but the final classification is
done using USDA triangle. Table 5 shows the different features of the different soil images. The
proposed system presented the feature vector in terms of a numeric value. This value is found
from the texture properties of the soil images, so a correlation is found between the feature
vector of the image and the texture of the soil. For that correlation, the final classifications of
the soil samples are done using multi support vector machine and kernel method.

A support vector machine is a supervised binary classification technique [21,27,28]. It is

proposed by Vapnik VN which is based on statistical learning theory [16,30]. The SVM is a very
popular tool in machine learning and always gives a better result in texture classification
[16,17,26,27], image retrieval and other classification problems [20,37]. Initially, it is used for
binary classification where classes are separated with the help of a hyperplane. The pixel
values nearest to this hyperplane are known as support pixel vector [19,22,27]. Later on, it is
extended for multi-class classification. In this paper, the multi SVM technique is used to classify
the images with the help of linear kernel function as defined by mercer’s theorem images
[14,15].

SOIL CLASSIFICATION USING MULTI SVM

For this research, 50 soil samples are collected and the textures are classified using the
hydrometer and USDA classification in the laboratory. The laboratory classification method

Table 3 – Number of Soil Samples per class with feature vector length.
Class Total of Feature VectorFeature VectorFeature Vector
Number length one singlelength per classlength per class
Sample image (Training) (Testing)
Clay Loam 7 1 × 193 4 × 193 3 × 193
Sandy Clay10 1 × 193 5 × 193 5 × 193
Loam
Silt Loam 11 1 × 193 6 × 193 5 × 193
 Clay 4 1 × 193 2 × 193 2 × 193
Fine Sand 2 1 × 193 1 × 193 1 × 193
Loam 2 1 × 193 1 × 193 1 × 193
Loam Fine sand 1 1 × 193 1 × 193 1 × 193
Loamy Sand 1 1 × 193 1 × 193 1 × 193
Sandy Clay 3 1 × 193 2 × 193 1 × 193
Sandy Loam 5 1 × 193 3 × 193 2 × 193
Silt Clay 1 1 × 193 1 × 193 1 × 193
Silt Clay Loam 3 1 × 193 2 × 193 1 × 193
Total 50 29 × 193 24 × 193
Bold defines the Total Class, Total numbers of soil samples and total size of Training Set.

gives 12 different classes of soils. It is found that more numbers of soil sample belong to
Clay loam, Sandy Clay Loam, and Silt Loam. Total 7 soil samples belong to Clay, 10 soil
samples belong to Sandy Clay Loam, and 11 soil samples belong to Silt Loam. In 2006,
Bhattacharya B and Solomatine [8] used three classes Support Vector Machine for the
classification of three soil classes and achieved 90.7% accuracy for the classification. In this
paper, initially, Support Vector Machine classifier is used to classify three soil classes such as
Clay loam, Sandy Clay Loam, and Silt Loam as these contain more number of soil samples. A
total of 15 images are used for training and 13 images are used for testing. In this approach, the
accuracy of the kernel based Multi Support Vector Machine is evaluated using cross-validation
and the algorithm of the accuracy is presented in algorithm 4.

Multi Class Support Vector Machine classifies the class of the given training vector according to
the given group and gives the result to which the class it belongs. Group in SVM is a vector
which can be numeric, categorical and also some- times it is logical. Each element in the group
specifies the corresponding row of the training. In this three-class classification, the size of the
training is 15x193 and group size is 1x15. The multi support vector machine classification is
implemented using Mat lab (R2017a). The training data and group data are used to create the
support vector machine classifier structure (SVM Struct). Again support vector machine
classifier structure (SVM Struct) and testing data are used to define the class to which the
testing sample is assigned. The group vector for three class classification contains the values
from 0 to 2 as the dataset contain only 3 soil classes. Now, based on the testing values, the
classifier classifies the class of the test image as it is assigned.

Twelve class soil classification using multi SVM

In this step, Support Vector Machine class classifier is tested for entire 12 classes of soil. The
size of the training is 29 × 193 and the group size is 1 × 29. The group vector con- tains the
values from 0 to 11 as the soil dataset contains 12 different soil classes. Now, based on the
testing values the
Algorithm 3: Algorithm for the result of soil classification (12 Classes)

Input: multisvm Result (test_result) Output: Classification Result

Step1: If (test_result == 0) then Class = Clay Loam.
Step 2: If (test_result == 1) then Class = Sandy Clay Loam Step 3: If
(test_result == 2) then Class = Silt Loam
Step 4: If (test_result == 3) then Class = Clay Step5: If (test_result == 4)
then Class = Fine Sand. Step 6: If (test_result == 5) then Class = Loam
Step 7: If (test_result == 6) then Class = Loam Fine Sand Step 8: If
(test_result == 7) then Class = Loamy Sand Step 9: If (test_result == 8) then
Class = Sandy Clay
Step 10: If (test_result == 9) then Class = Sandy Loam Step 11: If
(test_result == 10) then Class = Silt Clay
Step 12: If (test_result == 11) then Class = Silt Clay Loam
Accuracy for multiSVM

The soils are classified using SVM with a kernel function. A kernel is a similarity function and it
takes two inputs and spits out how similar they are [19]. In this paper SVM with a linear kernel
is applied to classify the soil classes. The reason of using linear kernel is presented below.

The linear kernel is a very simple and faster kernel approach of SVM. The linear kernel is good
when the images have too many features. This is because map- ping the data to a higher
dimensional space does not really improve the performance of the classifier [40]. In this paper,
5 different features of the soil images are extracted and it gives 193 different values of feature
for a single image. It is found that the feature size of a single image is large and it is 1x193. So,
the linear ker- nel with SVM is applied to classify the soil classes.

For this paper, non liner mappings of SVM are not per- formed for the dataset as it contains
many data points for the feature vector.

In this paper, data are trained using SVM with linear kernel. Linear kernel is faster and simple
kernel method. It requires optimizing the value of C parameter but others kernels require to
optimize the value C and c parameter. C in SVM is a regularization parameter which defines the
cost of misclassification. It defines the bias and variance of the model. The high values of C
mean low bias and high variance. The low values of C mean high bias and low variance. Low
bias and high variance always lead to over-fitting of the model whereas high bias and low
variance always lead to under-fitting of the model. In this paper, the value of C is considered as
1.0 and for this the model is neither over-fitted nor under-fitted.
The accuracy of the SVM classifier with the kernel is eval- uated using the holdout method of
cross-validation. During the time of accuracy calculation, holdout cross-validation of group
vector is performed and creates a classifier perfor- mance object, these accumulates the result of
the classifier for the testing vector. The algorithm for evaluating accuracy is presented by
algorithm 4. So for evaluating the accuracy of the system, the kernel function is trained for 500
iterations and 1000 iterations. But the maximum accuracy of the system is found for 500
iterations. To overcome the model over fitting, the samples are divided into training and testing
set as explained in Table 3 and the model is evaluated using cross- validation by setting the C
parameter of kernel as 1.0. Cross- validation is re-sampling technique used to evaluate the SVM
model as less numbers of samples are present in the soil dataset. The average testing accuracy of
the mode is more than 90% which is defined in Tables 5 and 6.

Algorithm 4: Algorithm for Accuracy of Classification

Input: Training Data, Group Data Output: Accuracy in percentage Step1: Consider iteration
= 500 Step 2: For i = 1 to iteration

Step 2.2: Take Train and load Group vector

Step 2.3: Perform Holdout Cross-Validation of Group Vector.

Step 2.4: Create Classifier performance object, CP for Group vector.

Step 2.5: Perform SVM training to create the SVM structure using Linear Kernel Function,
Training Vector and Group Vector.

Step 2.6: Now, Calculate the test_result by using Group Vector and Test Vector.

Step 2.7: Again Create Classifier performance object,

CP for Test vector by considering test_result and calculate the accuracy.

Step 2.8: Print the accuracy of the kernel in percentage.

For the three class classification, multi Support Vector Machine classifier gives an average of
95.72% of accuracy. The results obtained from the multi SVM classifier for the three classes are
presented in Table 5. Multi SVM gives good accuracy for the three class classifications because
the soil classes such as Clay Loam, Sandy Clay Loam, and Silt Loam have more soil
 Table 6 – Result of Multi-Class Classification (SVM).
Type Average Percentage of accuracy

Clay Loam 96.20%

Sandy Clay Loam 92.56%

Silt Loam 95.25%

Clay 90.54%

Fine Sand 92.58%

Loam 96.84%

Loam Fine sand 85.54%

Loamy Sand 81.25%

Sandy Clay 92.56%

Sandy Loam 95.65%

Silt Clay 84.25%

Silt Clay Loam 93.21%

Total 91.37%
Bold defines the total average accuracy of the 12 class classifier.

samples as compared to the other soil class. Different soil sam- ples are used for training and
testing and it is presented in Table 3. In the next step, Multi SVM is tested for the entire 12 soil
classes and the accuracy is presented in Table 6.

From Tables 5 and 6, it is found that the average accuracy of 12 class SVM classifier is less as
compared to the 3 class SVM classifier. It is because of the less numbers of soil sample in the
Loam Fine sand, Loamy sand, and Silt Clay soil. In 3 class classifier, the number of samples of
Clay Loam, Sandy Clay Loam, and Silt Loam are 7, 10, and 11 respectively. So, the variations
of soil samples are not much more to each others. But in 12 class classifier, only 1 sample is
present in Loam Fine sand, Loamy Sand, and Silt Clay respectively. So, the variations of these
three classes are more as compared to the Clay Loam, Sandy Clay Loam, and Sandy Loam,
Clay, Silt Clay Loam and Silt Loam soil. It leads to imbalanced data- set for these three classes.
But the dataset is not imbalanced with respect to the Fine Sand, Loam soil as these two classes
have only 2 soil samples. Though the samples of Loam Fine sand, Loamy sand, and Silt Clay
classes are less in number, the accuracy of the model is evaluated using the cross- validation by
setting the C parameter of kernel based SVM as 1.0 and achieved more than 80% accuracy for
the classes. As the accuracy of these three classes is less as compared to others, the overall
accuracy of the model is less from clas- sifying 3 to 12 classes.

Initially, three classes of soils are classified using three-class SVM classifier with linear kernel
and the accuracy of the classifier is evaluated using cross-validation. The accuracy of the
classifier is best for Clay loam soil and is 97.7%. But the aver- age accuracy of the classifier is
95.72%. Bhattacharya B and Solomatine [8] found an average of 90.7% accuracy for SVM class
classifier in case of sand, clay and peat type of soil. Wu W et al [27] found the 79.4%
accuracy for Clay soil, 99.2% accuracy for Loam soil, 66.1% accuracy for Sand soil.

The SVM classifier gives a good result for three classes of soil, so it is applied for all 12 classes
of soil. The results of the multi SVM classification for the 12 classes are presented in Table 6.
The classifier gives more than 90% accuracy for all types of soil class except the loamy fine
sand, loamy sand, and silty clay. Srunitha and Padmavathi [21] applied seven classes SVM
classifier for seven different types of soil samples and found 60.9% accuracy for correctly
classified soil seg- ments and 74.7% for correctly classified instances. Vibhute et al. [22] found
an average accuracy of 71.18% for five types of soil samples. They did not find the efficient
accuracy with SVM as due to the complex nature of the selected region and spectral information
of the soil samples. Guang Y et al. [30] achieved 96.67% accuracy for the seven class
classification using multi SVM with a linear kernel function. In this paper, the accuracy of the
multi-class SVM carries out with the help of linear kernel function with 500 iterations. The multi
SVM class classifier gives the best testing accuracy result for the Loam and Clay Loam soil and
it is respectively 96.84% and 96.2%. The average testing accuracy of the entire 12 class clas-
sifier is 91.37%. The accuracy for Loam Fine sand, Loamy sand, and Silt Clay are respectively
85.54%, 81.25%, and 84.25% and it is less as compare other soil classes because the same image
used as testing and training. The accuracy of the different machine learning method is compared
with the accuracy of the proposed approach and is presented in Table 7.

The proposed system is tested by collecting another 5 soil samples from the west zone of the
Guwahati with the help of rural farmers and then soil images are captured using the same
android mobile phone in front of the farmers. These 5 samples are labelled as Sample 1, Sample
2, Sample 3, Sample 4 and Sample 5. Out these 5 samples, sample no. 1 is Clay loam, Sample
no. 2 is Loam, Sample no. 3 and Sample no. 4 is Silt Loam and Sample no. 5 is Clay. The soil
samples are tested in the laboratory using the hydrometer method and their textures are
determined using USDA triangle. Then, the features of the soil samples are determined using
the pro- posed method and textured are classified with SVM. The pro- posed system gives the
same texture as the hydrometer method for those 5 samples and their accuracy is more than 90%
for each soil sample. Table 8 shows the comparative analysis of the hydrometer method and the
proposed method for that 5 soil class.

CHAPTER

CONCLUSION

In this paper, an image based texture analysis is presented to classify the soil images using multi
SVM and linear kernel function. The images are captured using android mobile phone camera
within West Guwahati Region. The three class classifier and multi-class classifier gives a good
performance for the real dataset except for the loamy fine sand, loamy sand, and silty clay.
Earlier, the texture of the soil is deter- mined with the traditional hydrometer method and USDA
tri- angle which is a very time and labour consuming process. A simple test of hydrometer takes
at least 24 h for the percent- age calculation of sand, silt, and clay. But the proposed sys- tem is
accurate and takes less time to classify the soil. It gives a fast and accurate result for soil
classification with the help of a support vector machine an android Smartphone.

REFERENCES

1.Gee GW and D Or. Particle size analysis. In: Dane JH and Top GC, editors. Methods of soil
analysis: SSSA, Madison, Wisconson; 2002. p. 255–293.

2.Bouyoucos GJ. Hydrometer method improved for making particle size analysis of soils. Agron
J 1962;54:464–5.

3.Jain AK, Farrokhnia F. Unsupervised Texture Segmentation Using Gabor Filters. J Pattern
Recognition 1991;24(12):1167–86.

4.Ahmed JA, Wesam MA. Image retrieval based on content using color feature. ISRN Comput
Graphics 2012:11.

5.Cristianini N, Shawe-Taylor J. Support Vector Machines. In An Introduction to Support Vector

Machines and Other Kernel- based Learning Methods. Cambridge: Cambridge University Press;
2000. p. 93–124.

6.Kecman V, Huang TM, Vogt M. Iterative single data algorithm for training kernel machines
from huge data sets: Theory and performance. In: Wang L, editor. Support vector machines:
Theory and applications. Studies in Fuzziness and Soft Computing, vol. 177. Berlin, Heidelberg:
Springer; 2005. p. 255–274.

7.Vapnik VN. The Nature of Statistical Learning Theory. New York: Springer; 1995.Pai PF.
System reliability forecasting by support vector machines with genetic algorithms. Mathemat
Comput Model 2006;43(3–4):262–74.

8.Soil Texture Calculator, NRCS Soils USDA. Link: https://www.

nrcs.usda.gov/wps/portal/nrcs/detail/soils/survey/?cid= nrcs142p2_054167.

9.Tian D, Zhao X, Shi Z. Support vector machine with mixture of kernels for image
classification. In: Proc. of the 7th Int’l Conf. on Intelligent Information Processing (ICIIP’12). p.
67–75.

Wang X, Yang H, Li Y. A new SVM-based active feedback scheme for image retrieval. Eng
Appl Artif Intell 2015;37:43–53Sun Y, Long Z, Jang PR, Plodinec MJ. Gabor wavelet image
analysis for soil texture classification. In: Proc. SPIE 5587, Nondestructive Sensing for Food
Safety, Quality, and Natural Resources. p. 254–61.

Bae SG, Yeon IK, Park SD, Kang CK, Zakaullah K. Effects of soil texture by soil addition on
the growth and quality of oriental melon (Cucumis melo L. var. makuw Mak.) under protected
cultivation. J Bio-Environ Con 2004;13(3):156–61.

Breul P, Gourves R. Field Soil Characterization: Approach Based on Texture Image Analysis. J
Geotech Geoenviron Eng 2006;132(1).

Zhang X, Younan NH, King RL. Soil texture classification using wavelet transform and maximum
likelihood approach. In: IGARSS 2003. 2003 IEEE International Geoscience and Remote Sensing
Symposium. Toulouse: Proceedings (IEEE Cat.No.03CH37477); 2003. p. 2888–90.

Zhang X, Vijayaraj V, Younan NH. Hyperspectral soil texture classification. In: IEEE Workshop
on Advances in Techniques for Analysis of Remotely Sensed Data, 2003, Greenbelt, MD, USA. p.
182–6.

Zhang X, Younan NH, O’Hara CG. Wavelet domain statistical hyperspectral soil texture
classification. IEEE Trans Geosci Remote Sens 2005;43(3):615–8.

Shenbagavalli R, Ramar K. Classification of soil textures based on Law’s features extracted from
preprocessing images on sequential and random windows. Bonfring Int J Adv Image Process
2011;1(1):15–8.
Bhattacharya B, Solomatine DP. Machine learning in soil classification. Neural Network.
2006;19(2):186–95.

Chung SO, Cho KH, Cho JW, Jung KY, Yamakawa T. Soil texture classification algorithm using
rgb characteristics of soil images. J. Fac. Agr. kyushu univ 2012;57(2):393–7.

Processing, Communication, Power and Embedded System (SCOPES). Paralakhemundi. p.

411–5.

Vibhute AD, Kale KV, Dhumal RK, Mehrotra SC. Soil type classification and mapping using
hyperspectral remote sensing data. In: Conference on Man and Machine Interfacing (MAMI).
Bhubaneswar. p. 1–4.

Haung J, Kumar SR, Mitra M, Zhu J, Zabih R. Image indexing using color Correlograms. In:
Proc. IEEE Computer Soc. Conf. on Computer Vision and Pattern Recognition. San Juan, Puerto
Rico, USA. p. 762–8.

Yu H, Li M, Zhang H, Feng J. Color texture moment for content based image retrieval. In: . Proc.
International Conference on image processing. Rochester, NY, USA. p. 929–32.Guwahati. Link:
https://goo.gl/maps/GDZ6mQFnqJt.

Zhao Z, Chow TL, Rees HW, Yang Q, Xing Z, Meng FR. Predict soil texture distributions using
an artificial neural network model. Comput. Electron. Agric. 2009;65:36–48.

Wu W, Li AD, He XH, Ma R, Liu HB, Lv JK. A comparison of support vector machines,

artificial neural network and classification tree for identifying soil texture classes in
southwest China. Computer Electron Agric 2018;144:86–93.

Kovacˇevic´ M, Bajat B, Gajic´ B. Soil type classification and estimation of soil properties
using support vector machines. Geoderma 2010;154:340–7.

Patel AK, Chatterjee S. Computer vision-based limestone rock-type classification using

probabilistic neural network. Geosci Front 2016;7:53–60.

Guang Y, Shujun Q, Pengfei C. Rock and soil classification using PLS-DA and SVM combined
with a laser-induced breakdown spectroscopy library. Plasma Sci Technol 2015;17 (8):656–63

Mistry Y, Ingole DT, Ingole MD. Content based image retrieval using hybrid features and various
distance metric. J Electro System Inform Technol 2017;5(3):874–88.

Vinayak V, Jindal S. CBIR system using color moment and color auto-correlogram with block
truncation coding. Int J Comput Appl 2017;161(9):1–7.
Mengistu AD, Alemayehu DM. Soil characterization and classification: a hybrid approach of
computer vision and sensor network. Int J Electrical Comput Eng 2018;8(2):989–95.

Huang T, Yang R, Hunag W, Qaio X. Detecting of sugarcane borer disesaes using support vector
machine. Informat. Proc. Agricult. 2018;5:74–82.

Chandan T, Thakur R. An intelligent model for indian soil classification using various machine
learning techniques. Int J Comput Eng Res 2018;8(9):33–41. Chung SO, Cho KH, Kong JW,
Sudduth KA, Jung KY. Soil texture classification algorithm using RGB characteristics of soil
images. IFAC Proc Vol 2010;43(26):34–8.

Kanimozhi T, Latha K. An integrated approach to region- based image retrieval using firefly
algorithm and support vector machine. Neurocomputing. 2015;151(3):1099–111.

Singh V, Misra A. Detection of plant leaf diseases using image segmentation and soft computing
techniques. Inf Process Agric 2017;4(1):41–9.

Barman U, Choudhury RD, Talukdar N, Deka P, Kalita I, Rahman N. Predication of soil pH

using HSI colour image processing and regression over Guwahati, Assam, India. J Appl Natural Sci
2018;10(2):805–9.

SVM:https://www.svm-tutorial.com/2014/10/svm-linear-kernel-good-text-classification/