MATLAB WORKSHOP

An Overview on Image Processing and Different Interfacings

Presenting you by:

Engr. Ghulam E Mustafa Abro

Faculty Member at Department of Electronics Engineering
Hamdard University Karachi
Contents
 Establishing Background of Image Processing
 MATLAB Basics
 Basic Codes of Image Processing
 GUI Tutorial & Exe. File generation
Introduction
 Engineers and researchers are always in the quest for the better, brief and
efficient methods to cope up the real world problems

 Hence there are several problems that can be solved using the techniques
of Digital Image Processing

 This Session at PSYWC 2016 suggests the importance and way out to solve
several issues with digital image processing (DIP) field
Digital Image Processing DIP
 First of all understanding DIP and why we implement DIP

 There are 2 big reasons for using DIP in general as mentioned

1. To enhance the pictorial information of an image &
2. For rendering an image more suitable for autonomous machine
perception
Image & Types
 An image is basically defined as two dimensional signal or a function
F(x, y) where x and y are the two coordinates horizontally and
vertically.
Image & Types
 An image has been classified into 3 types RGB, Gray Scale and Binary
Images

 RGB stands for Red, Green and Blue color intensity levels

 Whenever any image is captured initially it is basically consists of these

three color intensity levels or the combinational intensity of these colors

 Gray scale image means restricting the intensity levels in between 0-255
levels

 Whereas if these intensity levels are restricted or converted in between 0-1

intensity level showing either dark black or white color intensity levels,
such an image is known as Binary
Image & Types

Figure shows the types of images i.e. RGB, Gray Scale & Binary Images respectively
Formats of Image
Sr.no: Format Full Form Extensions
1. XWD X Window Dump .xwd
2. TIFF Tagged Image File Format .tif or .tiff
3. RAS Sun Raster .ras
4. PNM Portable Any Map .pnm
5. BMP Windows Bitmap .bmp
6. CUR Windows Cursor resources .cur
7. FITS Flexible Image transport System .fits or .fts
8. GIF Graphics Interchange Format .gif
9. ICO Icon resources .ico
10. JPEG Joint Photographic Experts Group .jpg or .jpeg

11. PBM Portable Bitmap .pbm

12. PGM Portable Grey Map .pgm
13. PBM Portable Bitmap .pbm
Classes of Images in MATLAB
Sr.no: Data type/classes Description

1. Single Single precision floating numbers ranges -1038 to +1038 (4

bytes/element)
2. Double Double precision floating numbers ranges -1038 to +1038 (8
bytes/element)
3. uint8 Unsigned integer 8- bit ranges in between [0 to 255] levels

4. uint16 Unsigned integer 16- bit ranges in between [0 to 65535] levels

5. uint32 Unsigned integer 32- bit ranges in between [0 to 4294967295]

levels
6. int8 Signed integer 8-bit ranges in between [-128 to 127] levels

7. int16 Signed integer 16- bit ranges in between [-32768 to 32767] levels

8. int32 Signed integer bit ranges in between [-2147483648 to

2147483647] levels
9. Char Character 2 bytes per element

10. Logical Binary values 0 or 1 (1 byte per element)

Real World Issues & FYP Suggestions
 Discussing real world issues that can be solved in a better way by using the
techniques of digital Image Processing are:
[1]. Fruit Grading and Healthiness checking
[2]. Efficient tool plaza System
[3]. Video processing based Manipulators
[4]. Color based Picking Robotic Arm
[5]. Alzheimer or Brain tumor Disease Detection
[6]. Facial Detection based Security Systems
[7]. Vein Recognition based Security Systems
[8]. Speech Recognition based Security Systems and etc.
Let‘s have a look on some
interesting videos
My First topic of Research
This whole procedure will be followed by a running conveyor belt, 2 AC
plungers and a wooden box in between them which will have 5 Mega pixel
camera mounted on it. The camera will be triggered by a brief algorithm of
digital image processing designed on the platform of MATLAB R2015a version
and conclude whether the fruit i.e. an apple is healthy or having some rotten
spots on its skin. Once results are shown, then algorithm will activate the
respective plunger for grading and sorting of apples.
 Major Techniques used before
Technique
K, C Means Method [1]
Wavelet Decomposition
method [2]
Ratio and Space indices
method [3]
Outline based Method [4]
Chain Code Method [5]
Gray Level co-occurrence [6]
Organization Feature
Parameter OFP [7]
Optimal Edged Based Shaped
method [11]
Gabor Filters [12]
Advantages Disadvantages
Provide the optimal results Includes salt and pepper
noise in results
Provides 94 percent efficient results Very expensive and needs
memory for computation
work
Detects fruit shape Having greater chances of
error
It is used for the detection of curve and Only used for outline
provides 93.2% efficient results detection
Need no memory for computational work and It never removes the noise
easy to implement from image
Used of the motion detection of an object and Very difficult to implement
provide high accuracy as well
Provide fine texture extraction of object But it detects Gaussian noise
This basically used to localize the object and Difficult to detect the aerial
shape object
This method provide the background extraction Implement on any one band
of any circle shape of colors
Our Technique & Methodology
 Here a solution has been demanded

 Having less disadvantages, maximum efficiency and an algorithm which can

be understood easily.

 In this regard the paper conveys the solution based on combination of

1. Gray Level Co-occurrence with
2. Threshold and segmentation technique.
Flow Chart
Results

Detection of Healthy Fruit (Manual)

Results

Detection of Unhealthy Fruit (Manual)

Results

Final Results (Manual)

Results
 Here are the results computed on MATLAB software based GUI

Result of Healthy Fruit

Results

Result of Unhealthy Fruit

References
 Fruit Detection with Multiple Features using Fuzzy Logic, Monika Sharma, Vibhuti Jaiswal and
Amit Goyal, International Journal of Advanced Research in Computer Science & Software
Engineering Vol. III issued on 11th November 2013. ISSN no: 2277128X.

 Method for fruits Recognition System by Woo Chaw Seng, Seyed hadi mirisaee, International
Conference on Electrical Engineering and Informatics 5-7 August 2009, Selangor, Malaysia.

 Image Processing for fruit shape and texture Fixture extraction- Review by trupen Meruliya, Parth
Dhameliya, Janish Patel, Dilav Panchal, Pooja Kadam and Sapan Naik, International Journal of
Computer Applications (0975-8887) Vol. 129 no.8 November 2015, Selangor, Malaysia.

 Corrado Costa & Francesca Antonucci & Federico Pallottino & Jacopo Aguzzi & Da-Wen Sun &
Paolo Menesatti. “Shape Analysis of Agricultural Products: A Review of Recent Research Advances
and Potential Application to Computer Vision” Springer Science Business Media, LLC 2011 (2011)
4:673-692.

 Sanket Regel, Rajendra Memane2, Mihir Phatak3, Parag Agarwal4 “2D geometric Shape and Color
Recognition using Digital Image Processing” IEEE Transcactions on Pattern Analysis and Machine
Intelligence, Vol. 23 No. 5, May 2001
References
 Mr. Hassan Sardar “Fruit Quality Estimation by color for Grading” Published in International
Journal of Modeling & Optimization, Vol.4, No.1 February 2014.

 Meysam Siyah Mansoory Isolating Healthy Banana’s from unhealthy ones based on feature, image
extraction, clustering methof and neural networks published Modern Sciences Vol.4, No.11,
November 2010 cited from www.ccsenet.org/mas

 H. Sardar, “Quality Analysis in grayscale color using visual appearance of a guava fruit”,
International Journal of Engineering Sciences, pp.46-56,2013.

 H. Sardar, “Quality Inspection and Grading Analysis of Fruit using Image Processing”,
International Journal of Agriculture and Food Science Technology, Vol. 3 no. 3, pp.176-179, 2012.

 H. Sardar, “A Role of Computer System for Comparative analysis using Image Processing to
promote agriculture Business”, International Journal of Engineering Research & Technology, Vol.
1, No. 9, 2012.
 Advice
“An apple in a day keeps doctor away”

Have a Break
Let‘s have SelfIEEE with
MATLAB