基于机器视觉的嫁接夹抓取点定位与朝向检测

针对现代农业自动化嫁接机的嫁接过程中，嫁接夹的分拣整理仍然需要使用刚性振动盘或人工参与，整体自动化程度有待提高的问题，提出基于机器视觉的嫁接夹轮廓链特征点定位方法与朝向检测；设计了一种自动化嫁接专用的嫁接夹，采用背光源打光方式，对每个阴影轮廓计算Hu矩，根据Hu矩匹配值进行分类以获得正确的嫁接夹个体轮廓；对每个嫁接夹个体轮廓采用轮廓链角分析方法，获得嫁接夹抓取点的像素位置；根据抓取点位置信息计算朝向角度；实验结果表明，嫁接夹抓取点平均定位误差为2.16个像素，嫁接夹朝向角度检测误差为0.40°；该方法可以准确地对嫁接夹进行视觉检测，对嫁接机的嫁接过程中，嫁接夹的自动分拣方面，具有机器视觉方面的价值。

Research on Gripping point Location and Orientation Detection of Grafting Clips base on Machine Vision

For the grafting process of modern agricultural automated grafting machine, grafting clips still requires the use of rigid shakers or manual involvement, and the overall automation degree needs to be improved, proposing a machine vision-based grafting clip gripping point positioning and orientation angle detection method; Designed a grafting clip for automated grafting, using backlighting method, calculating Hu moment for each shadow contour, and classifying according to Hu moment matching value to get the correct individual contour of grafting clip; Contour chain angle analysis is applied to each grafting clip individual contour to obtain the pixel position of the grafting clip gripping point; Calculate the orientation angle based on the gripping point position information; The experimental results show that the grafting clip average positioning error of grafting clip gripping point is 2.16 pixels, and the orientation angle detection error is 0.40 degrees; This method allows accurate visual inspection of grafting clips and is valuable in machine vision for the automatic sorting of grafting clips in the grafting process of grafting machines.