基于机器视觉的嫁接夹定位识别方法

Re search on Location Recognition Method of Grafting Clip Based on Machine Vision

为实现瓜科、茄科作物嫁接自动化上夹过程中嫁接夹抓取点及旋转角度的精确定位。通过基于HSV颜色空间的多值化提取算法,对机器视觉采集的目标图像进行处理,提取到目标的轮廓。采用基于轮廓链角的方法,识别轮廓的角点及质心。通过一系列计算,得出抓取点及旋转角度。结果显示,基于机器视觉的嫁接夹定位识别方法的抓取点判断,平均误差为2.1个像素;识别角点位置,平均误差为1.9个像素。本方法可应用于下一步的终端夹取操作。

In order to realized the precise positioning of the gripping point and the rotation angle of the grafting clip during the automatic clamping of the melon and solanaceous crops.The target image acquired by the machine vision was processed by the multi-valued extraction algorithm based on the HSV color space,and the contour of the target was extracted.Then,the extracted corner point and centroid of the contour were identified by the extracted contour angle based method.Through a series of calculations,the grab point and the angle of rotation were obtained.The results showed that the grabbing point of the grafting clip positioning recognition method based on machine vision was judged with an average error of 2.1 pixels;The corner position was recognized,and the average error was 1.9 pixels.The method could be effectively applied to the next terminal gripping operation.