摘要
为实现采茶设备的选择性采摘,针对嫩芽的识别与定位方法进行研究。通过对比各种色彩因子组合,提取GB分量对嫩芽图像进行灰度化处理,并在传统分水岭算法的基础上,引入BM3D去噪算法和灰度拉伸法对待分割嫩芽图像进行处理,分割出较完整的嫩芽图像。然后通过SURF算法进行特征点提取,结合RANSAC算法和极线约束去除误匹配点,在最小外接矩形法与双目测距原理的基础上完成茶叶采摘点的三维空间定位。最后通过采茶机械手进行定位实验,实验结果表明该方法基本满足采摘要求。
In order to realize the selective picking of tea picking equipment,the method of identifying and positioning the buds is studied.By comparing various color factor combinations,extracting GB components to grayscale the bud image,and on the basis of the traditional watershed algorithm,using the BM3D denoising algorithm and grayscale stretching method to process the segmented bud image.Segment a more complete shoot image.Then use the SURF algorithm to extract the feature points,use the RANSAC algorithm and the epipolar constraint to remove the mismatched points,and combine the minimum bounding rectangle method and the binocular distance measurement principle to complete the threedimensional location of the tea picking point.Finally,the positioning experiment was carried out by the tea picking manipulator,and the experimental results showed that the method basically met the picking requirements.
作者
姜宏涛
何博侠
章亚非
JIANG Hongtao;HE Boxia;ZHANG Yafei(School of Mechanical Engineering,Nanjing University of Science and Technology,Nanjing 210094,China)
出处
《机械与电子》
2021年第7期60-64,69,共6页
Machinery & Electronics
基金
国家自然科学基金资助项目(51575281)
中央高校基本科研业务费专项资金资助项目(30916011304)。
关键词
机械手
茶叶采摘
分水岭算法
视觉导引
manipulator
tea picking
watershed algorithm
visual guidance
