农业机械学报2017,Vol.48Issue(2):27-37,11.DOI:10.6041/j.issn.1000-1298.2017.02.004
基于机器视觉的定向播种用玉米种粒精选装置研究
Machine Vision-based Selection Machine of Corn Seed Used for Directional Seeding
摘要
Abstract
In order to meet the requirements for corn seed from directional seeding,a kind of corn seed dynamic selection machine was developed based on image processing technology.Firstly,the composition and working principle of this machine were introduced,and the device of blowing unqualified corn seed as well as its method of completing the blowing was designed.Also the dynamic detection method of corn seed images was developed.And then through the analysis of RGB color feature of these images,the extracting solution of the whole seed area and different color areas of seed in a corn seed image was obtained successively.Meanwhile,combined with the analysis of corn seed morphological characteristics,totally 20 detection indicators were set up to describe the eligibility of corn seed,such as perimeter,area,long axis,short axis.And the acceptable range of above indicators was determined through test statistics.In view of the above,the eligibility judgment methods of the following types of seed were analyzed respectively and executed successfully:seed with black embryo exposed in the tip,small seed,round seed,worm-eaten and damaged seed,moldy seed and other seed which did not conform to directional seeding.Furthermore,two points on contour line near two joints of adhesive seed were obtained,and it can be found that the ratio of the shorter distance of them along the contour line to the linear distance of them was larger than the corresponding value of any two points on the contour line of a single seed,which according to the adhesive seed can be detected.In the experiment,the results showed that the eligibility detection accuracy of corn seed was 96%,the judgment accuracy of adhesive seed was 99%,and the efficiency of blowing unqualified corn seed was 98%.关键词
玉米种粒/精选/图像处理/定向播种Key words
corn seed/selection/image processing/directional seeding分类
信息技术与安全科学引用本文复制引用
王侨,陈兵旗,朱德利,梁习卉子,代贺,陈洪密..基于机器视觉的定向播种用玉米种粒精选装置研究[J].农业机械学报,2017,48(2):27-37,11.基金项目
国家高技术研究发展计划(863计划)项目(2012AA10A501-5)、重庆青年人才计划项目(cstc2013kjrc-qnrc40001)、重庆市前沿与应用基础研究计划项目(cstc2013jcyjA80013)和重庆市教委科技计划项目(KJ1500321) (863计划)