This action might not be possible to undo. Are you sure you want to continue?
I hereby declare that project entitled “Online Classification of Rice Using Image Processing” is the work carried out at DU-2, Central Scientific Instruments Organization as requirement for the award of degree of Btech at A.C.E.T,Punjab Technical University under the guidance of Dr. H.K. Sardana.
Kamalpreet Kaur Certified that the above statement made by the student is correct to the best of our knowledge and belief.
HOD: col. Gurmukh singh
Dr. H.K.Sardana Scientist G CSIO, Chandigarh
Online Classification of Rice is developed for providing online service for the classification of rice grains using flatbed scanning and image processing. Here the neural based algorithm is used for classification of rice into 10 classes i.e. admixture, red, discolor, chalky, organic, small broken, big broken, Sound (Healthy), inorganic and damaged. The classification is based on physical parameter (Length, width, Area) along with color properties (Red, chalky, discolours). The procedure requires client to have a PC with Flat Bed Scanner and internet connection to the server configured to provide this service. Client has to upload rice image on the server and he will get the brief and detail report of rice classification including colour watershed image of rice sample which has indexing of rice grains. Detailed report (in zipped format) contains all physical and color features of each rice grain along with class it belongs to and is available for downloading. It yields the better accuracy than the more time consuming manual method. The developed system for rice classification takes less time in comparison to manual method.
Central Scientific Instruments Organization Sector 30 C, CHANDIGARH - 160030
CSIO is the foremost national laboratory for research, design and development of scientific instruments. It is one of the constituent laboratories of the Council of Scientific & Industrial Research (CSIR), which is administratively under The Central Scientific Instruments Organization (CSIO); Chandigarh is the Department of Scientific & Industrial Research of the Government of India. CSIR, India was constituted in 1942 as an autonomous body under the provision of the Registration of Societies Act Indo-Swiss Training Center and a housing colony. XXI of 1860. Situated in Sector-30C in Chandigarh, CSIO occupies an area of 120 acres. The CSIO campus comprises R & D laboratories,
CSIO presently employs 116 scientists, 108 technical officers, 413 scientific and technical supporting personnel and 156 administrative and other staff. It covers various disciplines like applied physics, bioengineering, industrial electronics, analytical instrumentation, digital & microprocessor based electronics, optics, electron optics, fiber optics, holography, electron & ion beam based instrumentation, metallurgy and mechanical engineering etc. CSIO will be completing its 50th year by 30th Oct, 2009. It covers various disciplines like: Applied physics Bioengineering Industrial electronics Analytical instrumentation Digital & microprocessor based electronics Optics, electron optics Fiber optics Holography Electron & ion beam based instrumentation,etc 1.1 HISTORICAL PERSPECTIVE CSIO was started in October 1959 in pursuance to the recommendations of a committee set up by the planning commission to formulate a scheme for the development of scientific instrument industry in India. Initially, it was located in the CSIR building at New Delhi. It moved to Chandigarh in 1962. An austere four-story building and the accompanying workshop were Inaugurated in December 1967 by the then President of India, Dr. Zakir Hussain. Another four-story block was added in 1976 for library, technical information and R & D activities. Indo-Swiss Training Center (I.S.T.C) was established at CSIO with Swiss assistance to meet the growing demands for well-trained instruments technologists. Dr. Fritz: Real, the President of the Swiss Foundation laid its corner stone in December 1962. Pandit Jawaharlal Nehru, the first 4
Prime Minister of INDIA, inaugurated the center in 1963. Shri M.C. Chagla opened the school building in December 1964, the then Union Education Minister. Shri Shivraj V. Patil, Vice-President, CSIR and minister of state inaugurated a 500 seating capacity auditorium for science and technology, Atomic Energy, Space, Electronics & Ocean Development in April 1985. 1.2 MAIN AREAS OF ACTIVITIES 1. Research, Design and Development of Scientific & Industrial Instruments, Components and Systems. 2. Service, Maintenance, Testing and Calibration of Instruments / Components. 3. Human Resource Development in the Area of Instrumentation. 4. Technical Assistance to Industry. 1.3 Decision Unit-2 DU-2 is one if the divisions of the CSIO. And all the projects in this division are under the supervision of the H.O.D. DR. H.K. Sardana (Scientist F) There are so many projects, which are being developed in this division. These are: • • • • • • Cephalometric Analysis Web Application for rice. HHSS (Hand held step scanner) for visual impaired persons. Electronic Portal Imaging Device(EPID) Fake currency detector. Real Time Image Processing Of Spackle in Fiber Optics Sensors.
providing more than one fifth of the calories consumed worldwide by the human species. It is the grain with the second highest worldwide production. Spain. USA and Japan. especially in Southeast Asia.g. size.2. after maize. Quality of edible products in general is based on a combination of subjective and objective factors. Rice is probably the most important grain with regards to human nutrition and caloric intake. Length (mm). Width (mm). shape and the number of broken rice kernels. Philippines and Australia) use different components of quality of agro products. we are considering the visual appearance and the related measurements.Its quality is based on a variety of properties such as the cooking texture. INTRODUCTION Rice is an important staple food for a large part of the world's human population. ] 6 . Length and width ratio are also independently (with very close margins) used to classify varieties of rice. Color other parameters are used to classify more than 20 varieties of rice. Whether a produce is acceptable for an intended use is determined by quality testing based on a fixed set of criteria. In the present phase of determining the rice quality. Chalk. Different countries (e. Rice Grain size and shape. color (whiteness and chalkiness).
i. requires trained personnel and results in a significant amount of incorrect classified rice kernels. The head rice yield. The many diverse uses of rice both domestically and for export. A sample layer of milled rice grains is placed on the sample holder which is placed on the glass plate of the scanner and covered with a black sheet of paper. Quality is based on a combination of subjective and objective factors.. The amount of broken rice kernels are specified when buying milled rice.1 Technique The aim of this research is to classify and grade rice grain sample according to FCI (Food Corporation of India) standard. FBS images provide uniform illumination independent of external light conditions. 2. require that quality to be evaluated according to its suitability for specific end uses. Manual inspection is very laborious. Transparency was used so that the glass of the flat bed scanner does not get damaged due to continuous use to place rice kernels.e.2. Broken rice kernels have normally only half of the value of whole or head rice. The length and width of rice kernels is generally measured using a sliding calliper one by one. The sample holder was made of transparency and a black sheet joined together at one end.2 Rice Quality There are many varieties of rice. The main distinction is between long and medium sized grains. The method developed for the determination of the color (whiteness and chalkiness). shape and size distribution of rice and the amount of broken rice kernels using Flatbed scanner (FBS) and image analysis software. size.The image will be acquired by PC through FBS. Whether rice is acceptable for an intended use is determined by quality testing 7 . the weight percentage of whole kernels remaining after milling is one of the most important physical characteristics that determines rice quality.
8 . environmental conditions. Plant breeders continually refine and improve genetic traits of new varieties required to produce the most desirable products. uniformity. Grain type categories are based upon physical qualities: length. storage. Image 2.No 1. and perimeter. shape. and processing techniques. Therefore physical properties such as size. Rice Admixture Description Mixture of lower quality rice. Rice quality is influenced by characteristics under genetic control. 2. S. characteristics are principally a function of handling.based on a fixed set of criteria. area. Organic Material which are eatable.3Milled Rice Classes Grain size and shape are among the first criteria of rice quality that breeders consider in developing new varieties. and distribution. In the latter case. The rice kernels appear in varying sizes and shapes. and general appearance are of utmost importance. width. The genetic makeup of a particular variety dictates to a large degree the grain quality characteristics. Rice is consumed as a whole grain.
Milled rice with length less than one quarter of the average length of the whole kernel. Inorganic 2. 9. 7. 4.4 Components 9 . Chalky kernels Which have characteristics similar to those of sound kernels. They are having standard length and width.g. 6. Red Red portion on the surface of rice. Discolour Yellowish colour on the surface of the rice. Sound Only these types of kernels have the required quality. e. Material which is not rice. 10. 8. Damaged Rice which is discoloured form the tip of the rice kernel.3. Small broken 5. or belly endosperm that affects their appearance and hence consumer acceptance. stones etc. Large broken Milled rice with length less than three quarters but more than one quarter of the average length of the whole kernel. back. but they have a large chalky portion in the centre.
that fulfil the prescribed specifications of the developing system software. Sample Holder 10 . shape.1. A standard flatbed scanner also called desktop scanner to be used to obtained images of the rice kernels. Flatbed scanner (FBS) 2. 3. Computer Under develop system is to be incorporated in any existing computer system. distribution of rice for the determination of the amount of broken rice kernels. FBS scanners are the most versatile and commonly used scanners found in all offices. Flat bed Scanner Flatbed scanner (FBS) will be used for acquiring rice sample image for the determination of the size.
11 . 3. Internet Connection: Internet Connection is needed to interact with the server. After acquisition of the image the file has to be uploaded to the web server using Internet Connection. Classification of rice. 5. 2. Feature calculation. Grading of rice.5 ARCHITECTURE OF THE SYSTEM a) Basic Steps: The basic steps in the classification and grading of rice are as follows: 1. Report Generation. 2. Upload the image. The base of the holder is transparent glass. 4.Sample holder is a container to hold rice sample. Scan the sample. 4. Sample layer of test rice sample is to be lain on this transparent area of holder for imaging. 6.
Wrapping up all the results and sending back to web server. 12 . Put the sample rice on the sample holder. Feature Extraction. Download the results.Grading of rice Input Sample Scanning Uploading Feature calculation Classification of rice Report Generation Grade of rice Classification and grading of rice basic steps b) Process to know the class and grade of rice: The steps involved in classification of rice are Put the sample holder on the scanner. Upload the acquired image. Watershed the image. Scan the sample. Binariesation of the image. Open the Internet Connection. Image will be sent to Application Server automatically. Report generation. Enter through authenticated login id.
It acquires image of the rice sample from the scanner and provides input to the image processing module. FBS are the most versatile and commonly used scanners. The interface between the scanner and PC is provided through USB port. It should be highly interactive in nature. In the designing of the system modules. It should be fully interactive. II. 13 . There are following modules in the project: I.c) Phases of the project According to the analysis of the requirement of the candidate system and the process. a system control. and at least one image control. this is to avoid the effects of reflection. A standard flat bed scanner also called desktop scanner is used to obtain images of the rice kernels The FBS is used with a sample holder. smoothing. Transparency was used so that the glass of the flat bed scanner does not get damaged due to continuous use to place rice kernels. image is acquired by scanner. Image processing Module: To perform image processing on an image. The User should be alerted with proper messages about the errors and the messages should be easily understandable and should guide the user right way. The black sheet used is dull. Image acquisition Module: In Image Acquisition Module. 3. The sample holder is made of transparency and a black sheet joined together at one end. The client/user can easily upload the image and download the results and user can also download its pervious results. the basic phases were decided for the system. watershed are performed on the image. these phases will be consulted. 4. 2. the website is to be designed. Client Server Communication Module: In this module. III. It should be User-Friendly and easily understandable. The first step in image processing is preprocessing the image which includes various processing operations like binariesing. you need an image processing control. The purpose behind designing of the User Interface is as follows: 1.
Watershed transformation is applied in conjunction with other processing operations to separate blobs from each other. Water shedding: After thresholding the aim is to separate the joined rice kernels if any. . Rice Grain on the Scanner (real Image) This process is done after scanning of image in order to easily differentiate rice grains from background. During water shedding color is extracted and in the image each blob is separated from other by removing border blobs. This operation is called threshold. For Binarization. thresholding is done in which one threshold value is set.IV. V. This is done by applying water shedding on the image. A minimum in the image is defined as a pixel or a set of connected pixels that is lower in value (or elevation) than all its neighboring pixels. All values above that threshold will be outputted as one (1) and values below threshold are outputted as zero (0). Binarization: A binarization method binaries an image by extracting lightness as a feature amount from the image. 14 . To identify the rice kernel in the FBS image. a binary image can be prepared by defining a range of brightness values in the original grey scale image as shown in figure below belonging to the foreground (rice kernels) and rejecting all of the other pixels to the background. It reduces an image to monochrome image (binary image).
as shown in the figure. Figure: Watershed Transformation image 15 .Watershed transformation. Compactness.Sound (%)). Feature Extraction Module Blobs are areas of touching pixels that are in the same logical pixel state. Feature extraction includes both Geometrical Features (like Area. hence that cannot be classified. Feature extraction is required for both classification as well as grading. these Border blobs are removed to separate the grains to extract red.Chalky (%). This watershed transform is then added with the original image to separate each rice kernels. Blobs are considered to consist of either zero or non-zero pixels. VI. When two or more blobs touch each other then it becomes difficult for the system to classify them because the system consider them as one blob due to which its Geometrical parameters increases. Watershed is able to separate only those blobs in which there is neck formation at the point of contact. depending on the foreground setting.Discolor (%).Damaged (%). This watershed image is anted with the gray scale image to remove border Blobs. Perimeter. is applied in conjunction with other. which may not lie in the normal range. while the alternate state is called the background state. Roughness. this is called the foreground state. Elongation) and Color Features(such as Red (%). The Figure highlights the watershed implemented to separate the joining blobs. green and blue color components. as shown in figure below. This problem is water shedding which needs to be overcome.
red. chalky. sound and inorganic. Classification at grain level Next step after pixel level classification is grain level classification.html file with all the classification details of the file with more interactive way. discolor. Creation of Report and make it available for the client Next step is to create a report in the form of . In geometrical features length. Output classes are admixture. In this. and red color has been considered. discolor (yellowish). classification is done on the basis of geometrical and color parameters. 16 . The range selection is done with the help of graphs obtained by the excel sheets of different samples of a particular class of rice and then taking into account the parameters on which that particular class depends. VIII.Figure: Touching blobs in a binary image Figure : After and Operation of watershed and original image VII. Wrapping up all the files used under processing with the output files as excel and binary images and upload it to web server and enable the availability to the client. small broken. breadth and area of each grain has been taken and in color features sound. chalky (whiteness). organic. big broken.
1 Overview of the existing system Existing System was a desktop application which can classify the rice using Rule Based Testing. and not applicable to provide services. This application has several restrictions like: It cannot be given to client for services because several backhand software’s need to have a licensed like Matrox Imaging. DRRI Hyderabad Not Known Introduction to the Existing System 3. Cuttak. 2. 1. TYPE End User Clients USER Mark fed. Under Rule based testing it can tell the results in between 50 to 60 percent correct.2.7 PARTIES ENVOLVED S NO. Hafed. CRRI. Some snapshots for the existing system are:- 17 . etc. It’s a It can’t work over the network. Government of India. Apeda.
Figure: Main Form for Existing System 18 .
Figure: Image Acquisition 19 .
There are changes that are needed to be made.2 Problem Assigned There are many features that are required to be added in the existing web application for rice classification. 3.Figure: Output for Existing System 3. 2. 20 . User can download the required result. These are:1. 4. Work on logout button. so that after logout back button get disable.server send massage to client if wrong image is uploaded. Error handling. Upload the image along with username and a token along with it for recognition.
4.Preliminary Survey and Feasibility Study An analysis model that is a part of the requirements phase is necessary as the first step for implementing the user requirements. • To make the web server.1 System Objectives The main objective of the project is to study the requirements of the user. 4. design a system and implement the system • To study the existing system and conduct the requirements study. requirement specification was prepared and the approval from the higher officials and approved by the project leader. One of the most important factors in system analysis is to develop good understanding of the system and its problems that enables the designers to identify the correct problems. leading to national ‘uniform procurement specifications and uniform testing’. suggest realistic solutions for them and also develop solutions to satisfy the users thus making the system acceptable in the organization. then determine the activities and procedures to be computerized.2 Scope The server based system with ‘knowledge and data’. To make available the end report to the user so that he can download it. • To enable the user to upload the image over client server architecture. 21 . Based on the observations made. can provide sample testing over internet. A systematic investigation of the system was carried out to determine the functions of the system and how they relate to each other and to any other system.
Availability of technically qualified staff in house for the duration of the project and subsequent maintenance phase. install. A feasibility study is conducted to obtain an overview of the problem and to roughly assess whether feasible solution exists prior to committing substantial resources to a project. It can calculate microscopic features of rice which cannot be distinguished visually by human eye very accurately. 1) Technical feasibility: can a solution be supported with existing technology? 2) Economical feasibility: is existing technology cost effective? 3) Operational feasibility: will the solution work in the organization if implemented? 4. Manual inspection is very laborious. more locations and more people can connect to the system.More frequent. it is generally desirable to conduct a Preliminary investigation called a feasibility study.3 Feasibility Study When complex problem and opportunities are to be defined. Organizations will need to make this assessment based on: • • Knowledge of current and emerging technological solutions. So a detailed study is carried out to check the workability of the system. 4. submit sample by employing their own scanner. 22 . Every project is feasible if given unlimited resource and infinite time. The primary objective of a feasibility study is to assess three types of feasibility. and effective set of resources.3. impact on the organization. computer and internet for certified report. operate and maintain the proposed system. requires trained personnel and results in a significant amount of incorrect classified rice.1 Technical Feasibility A systems development project may be regarded as technically feasibility or ‘practical’ if the organization has the necessary expertise and infrastructure to develop. ability to meet user needs. Precious time and money can be saved and untold professional embarrassment can be averted if an ill conceived system is recognized early in the definition phase. Feasibility study is undertaken to evaluate its workability.
4. including all the phases have been taken into account and it is strict minimum. the financial and/or technical capacity to procure appropriate infrastructure and expertise from outside. budget and staff resources invested during the design and implementation phase as well as infrastructure. 4. Operational feasibility: The system has been developed after extensive discussion with the end user and all the operational requirements has been taken into account during the planning and implementation stages. System Requirement 23 . training and maintenance costs incurred after implementation.3.3 Operational Feasibility A systems development project is likely to be operationally feasible if it meets the ‘needs’ and expectations of the organization. User acceptance is an important determinant of operational feasibility.• • Availability of infrastructure in house to support the development and maintenance of the proposed system. support. Hence the system is economically feasible. Hence the system is operationally feasible. The cost of developing the system. Where necessary.3. Hence it is technically feasible. Technical Feasibility: The proposed system can be easily developed using resources available in the organization. Economical feasibility: The proposed system can be easily developed using the resources available in the organization and they do not invest in procurement of additional hardware or software. The feasibility study of the proposed system has been carried out in all the three areas. These costs may include the time.2 Economic Feasibility A systems development project may be regarded as economically feasible or ‘good value’ to the organization if its anticipated benefits outweigh its estimated costs.
5.2 Software Requirements S. 5. such as functions. 3. Libraries used VC#. for windows by 24 . NO 1. organizing and manipulating data.The data is arranged in the rows and columns. 2. Excel Excel is an electronic spreadsheet program that can be used for storing. It has a number of built in features and tools.net Matrox Imaging Library (MIL) Neuro Solutions CGZIP Library Imaging eiStream. 1. NAME Database Tools used Development Language SOFTWARE Excel SQL SERVER Visual Basic. Scanner Type Maximum scan size Interface Optical Resolution Illumination System should have NAME Flatbed 11X9 inches USB 300 dpi Reflection 128MB RAM HARDWARE 5.NO.1 Hardware Requirements S. 2. 6.net 3.Net Asp. 4. It is a table which stores various types of data .
precision. charts. Matrox Imaging Library (MIL) Matrox Imaging Library (MIL) is a comprehensive collection of software tools for developing image analysis. processing. The key components of .formulas. display and archiving.Allows you to identify connected regions of pixels (blobs) within an image. analysis. simplified development and integration with a wide variety of programming languages. and data analysis tools that make it easier to work with large amounts of data. These operations include loading an image from file and transferring image data. Image processing control. or modify selected features of an image using Active MIL processing capabilities. machine vision.NET Framework. Visual Basic. The toolkit features interactive software and programming functions for image capture. Following is the list of Active MIL controls: Image Control: allows allocating and operating on images. and speed of image processing and analysis operations.NET Framework provides a managed execution environment. thereby reducing the time and effort required to bring your solution to market. formerly called Visual Basic .NET (VB. verifies the accuracy.NET Framework are the Common Language Runtime (manages the execution of code & provide different services like garbage collection etc. annotation. It consists of both systematic and random tests. 25 . These tools are designed to enhance productivity. Blob Analysis control. accentuate. and then calculate features of these blobs.Allows you to smooth. robustness.NET). is an object-oriented computer language based on the . The .) and . qualify.NET Framework Class Library. medical imaging and video analytics applications.Net Visual Basic (VB).
Allows you to set up the system on which to run an application. It also allows you to inquire about system-specific attributes. At the core of neural computation are the concepts of distributed. arcs. such as rectangles. such as the number of digitizers supported by the system. and is capable of generalizing to new. SQL Server is designed to store data in the central location (the server) and deliver it on demand to numerous other locations (the client). Some networks are supervised. lines. in that a human must determine what the network should learn from the data. the objective is to assign the input patterns to one of several categories or classes. so that they represent the probability of class membership. Neural networks perform computation in a very different way than conventional computers. Other networks are unsupervised. icrosoft SQL Server™ 2008 SQL Server is an SQL-compliant RDBMS. System Design 26 . usually represented by outputs restricted to lie in the range from 0 to 1. SQL is a command that allows us to modify or retrieve information from the database. where a single central processing unit sequentially dictates every piece of the action.Allows you to annotate or alter images with text. in that the way they organize information is hard-coded into their architecture. System control. previously unseen data. as well as basic graphics. A neural network is an adaptable system that can learn relationships through repeated presentation of data. In classification. Neuro Solutions Neuro Solutions is the premier neural network simulation environment. and dots. adaptive and nonlinear computing. It uses the backpropagation through time (BPTT) algorithm. SQL-compliant means that it uses the ANSI (American National Standard Institute) version of Structured Query Language or ‘SQL’.Graphic Context control.
1 FLOWCHART:- 6.2 E-R Diagram:- 27 .6.
net.3 High Level Design In the high level design.net in C# language. the program is divided into two parts. Client-Server communication is developed in asp. This provides the facility of image uploading and report downloading. This will analyze the uploaded image on server and will generate the corresponding output. user User Login address Process of Rice Classification Output User Database(Outpu t folder) 6.4 Data Flow Diagram 28 .username password Database(Sql table tbclient Image Upload Process email Contact no. 6. Image processing module is developed in VB.
net) as a background process. Executable program get the physical path of uploaded image as an argument and it will save classification result in . Client can download processed image and report. Online Classification of Rice using Image Processing Client-Server Communication at Web Server Image Processing at Image Processing at Application Server Application Server Level 0 Figure: High Level Design Overview 29 .Server runs the executable file of image analysis module (developed in VB.zip file format and colour watershed image on disk.
Request for Page Image Storage Requested Page Image to Upload Start process Temporary Images Image Processing and Report Generation Client Server Report Processed Image and Classification Result Classification result in format of zip folder Level 1 30 .
zip File Report Level 2 31 .5 What at server side going on DFD for Image Processing and Report Generation Figure: Data Flow Diagram Uploaded Image Image Storage Image PreProcessi ng Threshold Watershed Blob Geometrical Parameter Extraction Color Parameter Extraction Geometrical Feature Neural Color Feature based Classificatio n Report Generatio n .6.
8 Final Report Generation. 1. The project work is divided into two main modules with several sub modules are as following: 1. 1. 32 . 1.1 PROJECT DESCRIPTION “Web Application for Rice Classification” is developed for providing online service for the classification of rice grains. Client Server Communication Module 2.1 2.4 2. 2.9 Wrapping all the required files.5 2. Providing Login Page. 1. Application Server Side Module 1.System Implementation 7.4 Binarizing the Image. Controlling all processes by the Administrator.6 Registering the Client. Alteration in password and his submitted details.2 2.7 Feature Extraction into Excel. Administrator Login Page.3 2.2 Copying Image from Web Server.1 Triggering Image from Web Server 1.6 Water shedding the Image 1.3 Purge the duplicate files from the Web Server 1. 1. Upload and download Form.5 Smoothening the Image.
clients upload the image (scanned) and send to server and getting the results including excel sheet which classified the rice in different classes.bmp Before uploading . original image and . 7. If username and password were correct it goes to next form that is image upload otherwise a error message is appear on click of login button .1. > For Downloading the result click on DownloadResult Button.html format which is brief report including output chart.10 Saving the Zipped Output File to Web Server.bmp) . on the click of the upload button . watershed image. In this.2.After uploading the image name will be username+current date&time+abc. And there is a stored procedure Clientchecksp checks the password in database on click of login button. (2) Image Upload Form: In this form user upload the scanned image of rice sample.1 CLIENT SIDE INTERFACES (1) Login Form: In this form user has to fill his unique username and password and as we store the username and password in a database table name as tbclient. > For Previous Result Click on PreviousResult Button then Click on Download Button. This module works on client side. 33 . the Image would automatically get rename and get save in upload folder.2 Client-Server Communication Module: The GUI is designed using ASP. The new name of image will be (ex. If image name =abc.NET. 7.and the new name of imge would save automatically in database Image processing would start automatically(Application Server) then result would save in output folder of web server.
the rice classification of grains would be done using neural testing.(3) Download form: This form displays the results. Web Server) The next step is the processing of image. The preprocessing of the image includes various processing operations like binarizing. Whenever a file is encountered the file is copied to application server folder for processing. A binarizing operation 34 .html and . where the original image is saved after uploaded by the client and input folder at application server Figure: Input folder at Remote System (i. The first step for the application server is to create the Shared Input folder at Web Server. which is processed by backend processor.e. smoothing. 7. The results have many file like as colour – watershed image. .xls file becomes available in a single zip folder and user also get the previous result for downloading. The below images are of the Input folder of Web Server.3Application Server Side Module In the processing of the image. water shedding of image.
After performing binarizing the image is as shown below Figure: Binaries Image After binarizing the Watershed transformation is applied in conjunction with other processing operations to separate blobs from each other and also label the blobs as shown below: 35 .reduces an image to monochrome image (binary image). Binary images are useful when trying to differentiate the rice kernel from the background and to identify their geometric features.
Number of Grains and number of Grains in %. Weight in Gms (%).Also contain output chart as shown below: 36 .html file which contains file name. Information regarding various classes of rice on the basis of Weight in Gms .Figure: Labeled Watershed Image The final report of Rice Classification will be produced in the form of .
excel file. output chart. html final result. etc.Figure: Final Result All the files are saved at the output folder of the local system which includes the colored labeled watershed image. 39 .
The zip folder containing all required files is shown as below: Zipped files 40 .Figure: Output folder at Local System After completing the processing of original image all the necessary files and images wrapped in the zip folder.
Output Folder of Remote System So that the result is to be delivered to the client and client will download the output file as a zip file containing all desired results.After wrapping up the results the result is being saved to the output folder of remote system (i. Web Server).e. System Testing 41 .
and system levels of testing. In order to provide a through and rigorous examination of a software product. cases (inputs. development testing is typically organized into levels. the goals will be successfully achieved. As an example. A good test case has a high probability of exposing an error. System testing makes a logical assumption that if all parts of the system are correct. There is independence from coding.Software testing entails running software products under known conditions with defined inputs and documented outcomes that can be compared to their predefined expectations. procedures. requiring functions or modules that produce critical outcomes to be challenged with intensive testing of their fault tolerance features). A software testing process should be based on principles that foster effective examinations of a software product. and expected results). Applicable software testing tenets include: • • • • • • • The expected test outcome is predefined. Testers use different tools from coders Examining only the usual case is insufficient. it requires early planning in order to be effective and efficient. tools. outputs. documentation. A successful test is one that finds an error.). As such. and reporting criteria. Both application (user) and software (programming) expertise are employed.g. when correction will be extremely difficult. Another objective of testing is its utility as user oriented vehicle before implementation. Inadequate testing leads to errors that may not appear until months later. Test plans and test cases should be created as early in the software development process as feasible. critically. Testing Testing is a phase whose basic function is to detect the errors in the software.. Features to be Tested 42 . and imperfect activity. a software product’s testing can organized into unit. difficult. resources (personnel. methodologies. etc. environments. reliability. and/or safety issues (e. integration. It is a time consuming. Testing is done at different stages within the development phase. They should identify the schedules. The magnitude of effort to be applied throughout the testing process can be linked to complexity.
That is the modules at very bottom. Report Distribution. which has no subordinates. In regard to this project following modules are tested as follows: Image uploading When the client enters the website for testing his sampled rice. The image should fulfill the above requirements.1 Unit Testing This is the first level of testing. Image uploading.Following features are to be tested. 5. Image Processing 43 . Report Generation. 1. 8. 6. Feature Extraction. Qualifying Criteria and Data collection. he has to acquire an image from the flatbed scanner with the specified set of the dimensions of the image required i. 2. Efficiency of the algorithms used in the project Approach used for testing For this project Bottom Approach is used for testing. are tested. Behavior of the system under adverse conditions like the case where input to the system that is rice image is not acceptable to the system.e. 3. It is essentially for verification of the code for the project. Image Statistics. Then these modules are combined with higher-level modules for testing. 4. The image should be properly uploaded by the client to the web server. 2. at least 300dpi image without contrast stretching and must be in the reflective mode and 8-bit grayscale. TEST UNITS AND TYPES OF TESTS APPLIED Different parts for the fulfillment of the project are: 1. The response time of the algorithms used in the project 3. Image Processing. All the functional features specified in the requirement document are to be tested.
8. This testing is applied after the implementation of the system. So testing is required for this step. 8. So for proper understanding of the user this module is very necessary. Statistical analysis is done so that some features which are otherwise difficult to understand to the users can be easily explained. all the modules are integrated and tested to check for their performance. Qualifying Criteria and Data Collection This module is tested so as to analyze various statistical features such as excel generation. The system is tested after different varieties of rice sample are scanned. So testing of this module is very essential for the working of software. convex perimeter accurately and are stored popularly in the excel sheet or not .Testing is also required in this module as the number of rice blobs are also calculated in this module . the entire software is tested. It is tested whether 44 .3 SYSTEM TESTING During this. Feature Extraction This module is tested for whether it is able to calculate microscopic angle independent features such as length. This testing is applied to check the interfaces between the different modules. Testing is very essential so as to see that image is threshold at a right value. area.2 INTEGRATION TESTING After each module has been tested separately. This is very essential because on this value whole features which are to be extracted are dependent.If this module is not tested then our software will not work properly.Once the image of the sample is received from client the image is forwarded to the Application server for image processing operations such as Binaresation and Water shedding is applied. Image Statistics. For this purpose this module needs testing. Report Generation and Distribution This module is tested whether the image is processed and the requires output files are generated correctly and compress it in a zipped format and whether the result files are send to the right person or not. perimeter. The design is tested during this testing. breadth. So testing of this module is important because in this module analysis of color features is done.
the program code for the project was to be tested against the rules of the language.the system works properly or not according to the requirements of the system. the data fields and out-of-range items are tested.4 PROGRAM TESTING The program is tested for the two types of errors. It is tested whether the system generates the accurate grade or class of rice 8. Testing for the “Logical errors” play important role to check the correctness of the different modules 45 . And in Logical errors. A) Syntax errors B) Logic errors During checking the Syntax errors.
red. There may be only registering fee. Availability: System is expected to be available more frequently and more people can connect to the system by only submitting the sample having their own scanner. So we are providing a better alternative by introducing this project. 9. computer and internet. Machine vision system is relatively expensive.1 SOFTWARE SYSTEM ATTRIBUTES Reliability: This provides high degree of reliability. small broken (1/8th of sound). chalky.2 Comparison with Related Systems A high speed machine vision system was designed to sort rice into sound. big broken (3/4rth of sound). damaged. Cost: The cost of user is negligible as the system utilizes the standard scanner and computer. in which rice is used. discolored. admixture.Evaluation This part will evaluate the system in respect to the requirements specified and determine if it satisfies the goals specified in the project problem definition. Security: Only Authorized users after entering a valid username and password can access the site for rice testing. This prototype rice grain quality inspection system demonstrated high performance comparable to subjective human inspection making it more reliable. These methods use a CCD video camera with illumination source for image processing and analysis.3 Quality Factors This client-server application is very useful in commercial or any other areas. 9. organic and inorganic with an accuracy ranging from 87-95%. The system results in a significant amount of correct classified rice kernels. It has many qualities as follows: 46 . influenced by external light conditions and need an experience person to setup the system. 9.
that can provide sample testing over the internet. The performance of the system is dependent upon the sample placed by the user. There may only be licensing fee. The result is sample based and not upon the whole quantity. A server-based system with complete ‘knowledge and data’. Eight Visual Grading tests take more than 30 minutes of manual labour. computer and internet for a certified report. The long process of conducting a test manually does not allow enough ‘repetition’ of a test often leading to unreliable reporting The cost to the user is nil as the system utilises the standard scanner and computer. 4. 9. 47 . leading to national ‘uniform procurement specifications and uniform testing’ The same system can be customized to suit export requirements and BIS standards including annual relaxations applicable to all farmers during bad seasons. Transparency in the testing procedure among various public and private agencies More frequent. 3. Following are the risk factors for this project: 1.4 Risk Factors Risk factors are the factors that limit the solution of the problem or can course the system failure at later stages. submit sample by employing their own scanner. more locations and more people can connect to the system. The developed system does it in a few seconds. 2. Sample placement is time-consuming approximately 3 minutes. Care has to be taken that minimum numbers of grains are joined (less than 5%).
48 . the project is able to do the classification and report generation of sample rice image uploaded by client. Client has to upload rice image on the server and they will get the brief and detail report of rice classification including colored watershed image of rice sample which has numbering on each rice grain. The procedure requires client to have a PC with Flat Bed Scanner and network connection to the server configured to provide this service.10. Here classification of rice blobs is done on the basis of Neural based classification algorithm. This project has designed and implemented software to classify the grains of rice using image processing and neural testing At this point of time. Conclusions This part has provided an evaluation of the system developed in this project. Additionally ideas for future work have been provided that can be used for future development. The evaluation has tried to unveil both the advantages and disadvantages of the final system and shown how the different requirements specified have been fulfilled. Detailed report contains all physical and color features of each rice grain along with class it belongs to.
At this point of time. Work can be done for making user interfaces more interactive. As the number of simultaneous client requests to a given server increases. which becomes more accurate than Neural Network Training method.Future Enhancement Fuzzy Network Training method can also be used for this client server application. work can be done for opening and handling the client scanner by server according to requirement of process. Costing according to classes of rice. Traffic congestion on the network has been an issue since the inception of the client-server paradigm. the server can become severely overloaded 49 . client have to first scan the rice image and send it on the server for processing.
asp.asp?ID=4492 .36k http://www.aspx http://www.codeproject.net: Complete Reference 2) Web and URL’s: www.com/UploadFile/mahesh/DownloadUsingHTTP1213200501275 2AM/DownloadUsingHTTP.aspx http://dotnetslackers.com/Community/blogs/haissam/archive/2007/02/02/RunExecutable-file-in-ASP.codeproject.com/KB/aspnet/simpleuploadimage.sourcecodesworld.net/QuickStart/aspnet/ .com/kb/318597 www.freevbcode.asp.50k – www.com/ShowCode.References 1) Books: VB.aspx www.com/ http://support.NET.com 50 .net: Complete Reference HTML: Complete Reference C#.microsoft.36k www.net/QuickStart/aspnet/ .csharpcorner.
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue reading from where you left off, or restart the preview.