A study on Anti-Collision Path Planning Method of Grape Picking Based on Binocular Vision
At present,the mechanized and artificial grape harvesting methods put into actual production have some problems such as poor adaptability,high cost,and large labor gap.It is necessary to develop intelligent picking equipment with anthropomorphic operation to improve efficiency.Intelligent path planning is an impor-tant research content for realizing non-destructive grape picking.Based on binocular stereo vision,the grape anti-collision picking path planning method was studied for the trellised grape clusters in Ningxia.The binocu-lar grape image was corrected by the polar line correction method,and the grape cluster target area of the left eye image was divided based on the combination of multiple color space models.The parallax value between the left eye segmentation image and the right eye original image was calculated by the local stereo matching algorithm,and the three-dimensional coordinates of all pixels within the grape cluster area were solved to form a collision free path.The results show that the proposed method has small data processing capacity and accu-rate calculation results,which can provide a theoretical basis for the development of intelligent grape picking equipment.
NETL
NSTL
万方数据
