玉米种子实时检测分选装置设计与试验

Design and Experiment of Real-time Detection and Sorting Device for Maize Seeds

摘要:

针对我国玉米种子人工分选效率低、错分率高、缺少自动检测分选装置等问题，设计了一种玉米种子实时检测分选装置。该装置由进料单元、检测单元、分选单元和控制系统组成。下位机采用MSP430，与上位机实时通信，并控制分选执行机构，上位机采用Matlab 2014b软件对玉米种子图像进行实时处理，并输出识别结果。为了便于采集玉米种子图像，设计了种子分离机构。根据霉变玉米种子与正常玉米种子表面颜色的差异，设计了一种基于HSV颜色空间划分的玉米种子识别算法，并提出了一种玉米种子排序策略，实现了玉米种子的精确分选。该装置对单幅图像的采集和处理时间约为0.7s，分选速率最高为680粒/min，霉变玉米种子识别准确率为100%，裝置总体分选准确率不低于94%。该装置实现了从玉米种子进料到分选的全自动化，能够对霉变玉米种子进行实时检测和分选。

Abstract:

In view of the problems of low efficiency of artificial separation, high rate of wrong separation and lack of automatic detection and separation device of mildewed corn seeds in China, a real-time detection and separation device was designed. The device consisted of a feeding unit, a detection unit, a sorting unit and a control system. The MSP430 was used to communicate by lower computer with the upper computer in real time and control the sorting actuator. The Matlab 2014b software was used by upper computer to process the corn seed image in real time and output the recognition results. In order to facilitate the image acquisition of corn seeds, a seed separation mechanism was designed. Based on the difference of surface color between mildewed corn seeds and normal corn seeds, an image algorithm for identifying mildewed corn seeds simultaneously in a single image was designed based on HSV spatial color division of image. The algorithm can realize pixel level mildewed corn seed recognition, and a sort strategy of corn seeds was proposed to achieve accurate sorting. The experiment showed that the device can realize the real-time detection and sorting of mildewed corn seeds. The acquisition and processing time of a single image was about 0.7s, the highest sorting speed was 680 grains per minute, the recognition accuracy of mildewed corn seeds was 100%, and the overall sorting accuracy was no less than 94%. The device realized the whole process of corn from feeding to sorting automatically, and can achieve the purpose of real-time detection and sorting of moldy corn seeds. The device adopted machine vision technology to carry out non-destructive detection of corn seeds, which provided technical support for the related research and development of later corn seed grading system.