The Software Design And Implementation Of Crops Sorting System Based On Machine Vision

With the continuous improvement of social productivity,machine vision technology has been introduced into a variety of industrial and agricultural production.Color sorting machine is one of the applications for high-speed,accurate sorting of the target material,which greatly reduces the labor input,while improving production efficiency.This thesis is focused on the color sorting machine's machine vision algorithm part and the designation of the control system program.Firstly,we analyzed the current development status of sorting technology,put forward and implemented a software and hardware architecture scheme based on machine vision for crops sorting system according to the actual production.This thesis mainly focused on the software design and implementation process of the system.Then we built an image processing platform based on TMS320C6748,analyzed and selected the appropriate main control processor,image sensor and system transmission bus in detail,and established the system control software platform.Then,the material sorting algorithm was designed,including filtering algorithm,waveform correction algorithm,color space processing,material identification algorithm and material location algorithm.The first three algorithms were used to preprocess the image,improve the clarity and stability of the image,and the last two algorithms were the core of the color sorting algorithm,which was used to identify and locate the material.The former algorithms greatly affected the accuracy of the latter.Then we designed the program of control system based on MFC,built the main framework,which contained the machine detection,program management,machine lock-code management function modules and so on,of which the most important was the automatic color selection function module,including material analysis algorithm and pixel value clustering analysis algorithm.The material analysis algorithm was used to extract the material selected manually from the target image,and the pixel value clustering analysis algorithm was used to analyze the pixels value and eigenvalues of the extracted material.After a year's development and debugging,the color sorting system designed in this topic realizes all functions,can high speed,accurately identify and remove the target crop,and at the same time the system control procedure provided the perfect and friendly man-machine interactive function.The user can be familiar with the system in a short time,the grains production efficiency has been greatly improved.