基于机器视觉的联合收获机作业路径规划方法

Path Planning Method for Combine Harvester Operations Based on Machine Vision

研究了基于机器视觉的联合收获机作业路径规划方法,解决联合收获机作业环境中因障碍物过多影响作业效率的问题。利用双目视觉相机采集联合收获机作业图像,选取九点标定法进行双目视觉相机的手眼标定。采用Canny算法利用高斯滤波器平滑处理图像,计算图像的梯度方向与幅值,利用双阈值连接联合收获机作业图像边缘,检测联合收获机作业图像边缘,划分收获目标与障碍物。依据人工势场法原理,利用联合收获机作业的环境信息构建一个势场。合成势场内联合收获机受到的引力与斥力,依据联合收获机的移动速度及合力方向,输出联合收获机的作业路径规划结果。试验结果表明,该方法可以依据联合收获机的作业环境,规划作业路径,有效规避障碍物,适应性好。

Path planning method for combine harvester operations based on machine vision was studied to address issue of excessive obstacles affecting operational efficiency in combine harvester operating environment.Binocular vision camera was used to collect im-ages of combine harvester operations,and nine point calibration method was selected for hand eye calibration of binocular vision cam-era.Canny algorithm was used to smooth image using a Gaussian filter,calculate gradient direction and amplitude of image,connect edges of combine operation image with dual thresholds,detect edges of combine operation image,and divide harvesting target and obstacles.Based on principle of artificial potential field method,a potential field was constructed using environmental information of combine harvester operation.Gravitational and repulsive forces received by combine harvester in synthetic potential field,based on movement speed and resultant direction of combine harvester,operation path planning results of combine harvester were output.Experi-mental results showed that,this method could plan operation path based on working environment of combine harvester,and planning results of operation path effectively avoid obstacles,with high applicability.