摘要
为实现油炸花生实时检测和拣选,设计一套基于机器视觉的智能分拣系统。针对系统功能,搭建传输模块和图像采集模块框架,根据HSV颜色模型的明度通道图像进行二值化和边缘轮廓提取,计算目标的饱满度、褶皱度和局部区域灰度差特征,组成多参数组合分类器判别油炸花生类别。通过文中提出的手眼标定方法,将缺陷粒坐标信息转换到机器人坐标系,控制器采用微分拦截算法完成拣选过程。试验证明,拣选系统对已知类别的100粒目标,检测准确率达95%以上,机器人3次重复定位精度1 mm以内,抓取准确率近94.7%,具有良好的工业实用性,并保证了较高精度。
In order to realize the real-time detection and sorting of fried peanut,an intelligent sorting system based on machine vision is designed.According to the system function,the frame of transmission module and image acquisition module is built.According to the brightness channel image of HSV color model,binarization and edge contour extraction are carried out.The plumpness,wrinkle degree and local gray difference of the target are calculated,and a multi parameter combination classifier is formed to identify the fried peanut category.Through the hand eye calibration method proposed in this paper,the defect particle coordinate information is transformed into the robot coordinate system,and the differential tracking algorithm is used to complete the picking process.The experiments show that the accuracy of the sorting system is more than 95%for 100 targets with known classification.The robot positioning accuracy is less than 1 mm for three times,and the grasping accuracy is nearly 94.7%.It has good industrial practicability and ensures high accuracy.
作者
谷林峰
李亚
GU Linfeng;LI Ya(Tianjin Key Laboratory of Integrated Design and On-line Monitoring for Light Industry and Food Machinery and Equipment,College of Mechanical Engineering,Tianjin University of Science&Technology,Tianjin 300222)
出处
《食品工业》
CAS
2021年第8期188-192,共5页
The Food Industry
基金
天津市科技计划项目(18ZXRHGX00020)
天津市企业科技特派员项目(19JCTPJC52100,20YDTPJC00130,20YDTPJC00140)。
关键词
机器视觉
油炸花生
分拣
缺陷检测
手眼标定
machine vision
fried peanut
defect detection
grab
hand-eye calibration
