Study On The Pruning Point Locating Method Of The Grapevine Winter Pruning Robot

Grapevine pruning is a key procedure to balance wood production and fruit, to promote the high and stable yield. The current pruning is still manual work mainly. The mechanization level is low and the labor is intensive so that the growers have a strong demand for more efficient and labor-saving devices. Thus researches on automatic and intelligent devices for grapevine winter pruning are of great importance.In this paper, the three-year-old fence Y type grapevine was taken as experimental object. Aiming at the greenhouse cultivation environment, the grapevine winter pruning operation system was designed. According to the image acquisition environment of the system, the buds detection and pruning point location were researched under the lab environment, laying the foundation for the further development of grapevine winter pruning robot system. The contents and results of the research can be briefly summarized as follows:(1) A kind of winter pruning robot system was designed for the fence Y style grapevine with more than three years, including automatic navigation car platform、image acquirement box、form pruning module、pruning module. By automatic navigation walk, the system could realize pre-pruning. Then winter pruning operation could be completed with machine vision.(2) The image acquisition platform was built in the lab environment. The grapevine color image was captured using the CCD monocular camera. B channel was chosen to transform color image to gray image. A binary image was gained after the image preprocessing such as filter、threshold segmentation and close.(3) Parallel thinning algorithm was used to extracting the skeleton of the grapevine. Comparing the application effects of the morphology thinning algorithm、Zhang thinning algorithm and Rosenfeld thinning algorithm, Zhang thinning algorithm was selected because for extracting the skeleton of grapevine image.(4) Harris algorithm was selected to detect the buds on the grapevine skeleton image.115different acquisition angles of the experimental sample image were detected. As a result, the buds average correct detection rate was76.5%, the buds average recognition rate was66.4%. With the identified buds information, a kind of pruning point positioning algorithm of the grapevine was proposed. The experiment result showed, the algorithm has a certain accuracy and feasibility(5) On Windows platform, feature point detection system for the sample grapevine winter pruning was developed by VC++6.0. The system was used as research and experiment tool for the grapevine pruning related algorithm.