摘要
应用机器视觉的识别和定位方法,设计了一套系统能够进行自动抓取、翻面、旋转操作,最终将陶瓷基片以统一位姿放到指定位置。从视觉系统采集的原始图像经过直方图均衡化、最大熵阈值分割、Sobel算子提取目标边缘后获得边缘梯度。使用基于边界的模板匹配算法搜索图像中的目标,将识别到的目标应用一阶矩计算质心坐标,二阶矩计算旋转角度。获得陶瓷基片位置和角度信息后传递给计算机,计算机触发控制系统驱动执行系统拾取陶瓷基片。为了验证系统的可靠性和实用性,在不同遮挡程度下对陶瓷基片进行识别,实验表明系统对遮挡较小目标具有一定识别率。计算目标实际角度和匹配角度的平均误差约为0.58,拾放的准确率为99%,具有一定的实际应用价值。
The recognition and position of machine vishion is applied,which is capable of automatically grabbing,turning,rotating ceramic substrate and finally placing it into the designated location in an uniform way.Images acquired from the visual system will go through histogram equalization,maximum entropy threshold segment and sobel edge detection to extract the edge of the ceramic substrate.Using the boundary-based template matching algorithm to search targets in the image.The first moment can be used to calculate its centroid coordinate,and the second moment can be used to calculate its rotation angle.The position and angle information is transmitted to the computer,so the execution system can be driven by the control system to pick up the ceramic substrate.In order to verify the reliability and practicability of the system,the ceramic substrate is recognized under different degrees of occlusion.
出处
《工业控制计算机》
2019年第3期95-97,共3页
Industrial Control Computer
关键词
通信工程
分拣系统
机器视觉
模板匹配
矩
communication engineering
sorting system
machine vision
template matching
moment