关键词: 枝叶修剪, 视觉伺服, 神经网络, 图像处理

Abstract: In order to solve the problem that the tomato branch and leaf pruning manipulator could not accurately find the pruning points of the tomato branch and leaf pruning manipulator,  and to ensure the efficiency of the tomato branch and leaf pruning manipulator,  a tomato branch feature point recognition method based on different edge types was proposed. Based on the neural network optimized by the particle swarm optimization algorithm (PSO-BP),  in this paper,  an “eye on hand” visual servo control system was designed. By dividing the horizontal and vertical edges of tomato plants,  the horizontal and vertical edges were extracted,  and the short edges of tomato leaves were removed by threshold segmentation. The horizontal and vertical coordinates of the intersection points were recorded. The intersection points were restricted to the upper edge and lower edge of the lateral edges to obtain the dual branch points of the tomato. The test results showed that the designed visual servo system could successfully identify the tomato side branch points. The simulation test showed that the error between the actual value and the expected value of the image feature was between 3-5 pixels,  and the fivedegreeoffreedom manipulator moved within 4 seconds,  which provides a reference for further research on agricultural pruning robots.

Key words: branch and leaf pruning, visual serving, neural network, image processing