基于机器视觉的马铃薯智能分级系统  

Intelligent potato grading system based on computer vision

作　　者：邓立苗[1] 韩仲志[1] 徐艳[1] 熊凯[1] 

机构地区：[1]青岛农业大学理学与信息科学学院,山东青岛266109

出　　处：《食品与机械》2014年第5期144-146,共3页Food and Machinery

基　　金：国家自然科学基金(编号:31201133);青岛市科技发展计划(编号:11-2-3-20-nsh)

摘　　要：为实现马铃薯外观品质自动分级,构建马铃薯智能分级装置,提出一种简单易行的马铃薯智能分级控制方法和流程,并基于Visual C++环境实现马铃薯智能分级软件系统。该系统通过下位机发送信号给上位机来控制摄像头拍照,上位机对所拍摄的马铃薯图像进行处理与分级检测,并将检测结果发送给下位机智能分级装置,最后分级执行器执行分级。经测试:该分级控制方法简单易行且运行稳定,不受机器运行速度的影响,基本消除了强电干扰所构建的分级系统分级结果稳定,分级精度达到90%,可满足马铃薯实时分级的要求。In order to realize automatic grading of potatoes by external appearance, the intelligent potatoes grading device was constructed based on computer visual. A simple and feasible grading controlling method was gotten, and extraction algorithms of grading features were presented. And then the software system was implemented based on Visual C ＋ ＋ integrated development environment. The process of the system is as followed： The lower computer controls the camera takes photos by sending signal to the visual inspection system on the upper computer. And then the upper computer detects the potato from three aspects of shape, color and defects. The detection re- sult of the upper computer will be sent to the lower computer by seri al ports and then control the grading actuators to pick up the defects. Results showed the grading method described above was steady, and its precision could be up to 90%, which meet the needs of the real time detection.

关 键 词：分级系统 机器视觉 检测 传感器 

分 类 号：TS215[轻工技术与工程—粮食、油脂及植物蛋白工程] TS203[轻工技术与工程—食品科学与工程]