关键词: 枇杷枝条, 图像识别, 机器人视觉, 框架提取

Abstract: Proposed in this paper is a method of image recognition and framework extraction for loquat branches，which is based on the analysis of the image characteristics of loquat branches． In this method，first，a common seg-mentation threshold is determined by transforming the brightness intensity． Then，the initial characteristic image of branches is obtained by the superposition of the segmented image and the converted binary image． Moreover，the linked feature of loquat branches is used to extract the image feature again and to eliminate the background noise．Finally，a compensation of the boundary contour and non-rational discontinuity of branches image is made to imple-ment the extraction of branches frame and to acquire the branche diameter and the center coordinates for robot prun-ing． A case study is also carried out to verify the accuracy of the proposed method，and an identification rate of 91．2% is obtained．

Key words: loquat branch, image recognition, robot vision, framework extraction