基于机器视觉的药片表面缺陷识别与分拣系统设计

摘要

针对制药企业采用人眼识别药片表面缺陷受光线强度及视觉疲劳影响,存在漏检和低效等问题,利用机器视觉技术,进行了药片图像预处理,设计了药片缺陷检测及分类、异色处理算法并处理;在此基础上,利用反问题的最小二乘解算法确定缺陷药片的物理位置,控制3自由度 Delta机器人实现缺陷药片的分拣.经过测试,系统分拣成功率高,且易于实施,达到了设计要求.%Aiming at problem that pharmaceutical companies use human eye to identify tablet surface defects which cause visual fatigue affected by light intensity,undetected and inefficient problems exist,use machine vision technology for tablet image pre-processing,tablet defect detection and classification of different color processing algorithms and processing;on this basis,the least squares solution of inverse problems for the use of algorithms to determine the physical location of the defect tablets,3 DOF Delta robot control pills sort of defect.By testing,the system has high sorting success rate,and easy to implement,and meet design requirements.