Research And Development Of Fruits And Vegetables Traceability Scale System Based On Image Processing

As the government paying more and more attention to the construction of food traceability system, the Minsitry of Commerce is carring out the project known as "meat and vegetables traceablity system" in full swing. Traceability scale is a terminal equipment which contains functions like weigh products, print bills, wireless communication, settle accounts, upload data and so on. Connected by the internet of things technology, traceablity scale has been a key part of food traceality system. The traditional traceablity scale mainly input fruits and vegetables’ traceability information by manual button input or scanning two-dimensional code lable or RFID label pasted on the product. These methods not only requires a lot of manpower and material costs, but also is not conductive to the automatic selling fruits and vegetables. In order to solve the above problems, this paper researches and develops the fruits and vegetables traceability scale system based on image processing. The system uses the C/S architecture model, and connects client and server by wireless network for data communication. The client runs on an embedded system platform, responsible for sales of fruts and vegetables and image recognition. The server runs on a PC platform, is responsible for the management of client sales information.The main contents and innovations of this paper are as follows:(1) The overall scheme of fruits and vegetables scale traceability system is designed. The system uses the C/S structure, through the wireless local area network, server software can remotely manage client data. The needed hardware and software development environment is built, which lays the foundation for the development of traceability scale system.(2) In order to construct fruits and vegetables image database, this paper develops an image acquisition device and uses it to take products images. The image database contains 13 kinds of fruits and vegetables images under 6 kinds of light conditions to simulate the real sale environment as far as possible. Then, image preprocess alogorithm is researched, and proposes a kind of K-means clustering image segmentation algorithm based on Lab color space. Comparing test with the classical segmentation alogorithms shows this algorithm can segment fruits and vegetables image more accurately with good illumination robustness.(3)The global color histogram feature and local binary pattern is studied in depth, and OAO, OAA and DAG three kinds of multiclass SVM alogorithm are compared in detail. On this basis, proposes a multi-class SVM fruits and vegetables recognition alogorithm based on color and texture features. The comparison recognition experiments based on the fruits and vegetables database, show that the proposed GCH+LBP+OAO SVM algorithm for fruits and vegetables recognition has better recognition ability and practicability.(4)The software development of client and server of the traceability scale system is completed. The client software development includes the development of image acquisition module based on V4L2, the development of weighing data serial communication module and the Linux Wi-Fi communication module development based on wpa_supplicant. During the development of server software, the database of traceability scale system is firstly detailed designed, then the information management UI is developed. Finally, the communication protocol between the server and the client is defined and the socket communication programming based on Tcp protocol is realized.(5)For the purpose of testing the function of the whole system, a field joint test envirionment is built. Then the fruits and vegetables traceability scale system based on image processing is tested in all important aspects. In the test, the system runs well and the functions are realized right, and the test results show that this system has high practical value.