自主行走拖拉机道路识别与路径导航——基于激光扫描测距

Road Identification and Path Guidance for Autonomous Walking Tractor Based on Laser Scanning Range Finding

针对拖拉机自主行走导航的应用需求,提出了一种应用于拖拉机自主行走的田间道路边缘检测与导航算法。该方法利用激光测距仪获取当前视场内路面、作物及障碍物信息,有效区分作物和道路,从而使拖拉机沿着田垄间行走;利用基于Sobel算子的机器视觉来获取导航航向角度,通过调整左右轮不同的速度,最终实现航向的驱动。为了验证该方案的可行性,在理想的实验室状态下,对激光扫描序列进行了测试,由激光扫描序列的波动性,验证了其对道路和农作物区分的准确性,从而验证了方案的可行性。由拖拉机自主导航实验发现:拖拉机可以准确地完成单垄行走和复杂多垄间行走的自主导航,最大的导航误差仅为0.033m,比传统的导航算法精确性更高,从而验证了方案的可靠性。

Aiming at the application demand of tractor autonomous navigation,a road edge detection and navigation algorithm for tractor autonomous walking is proposed.In order to obtain the current pavement,crops and obstacles in field information using laser rangefinder,the method effectively distinguish between crop and road,so that along the ridge between the walking tractor; the use of machine vision based on Sobel operator to get the navigation course angle,by adjusting the theory about different speed,and ultimately drive course.In order to verify the feasibility of this scheme,the ideal state in the laboratory,the laser scanning sequences were tested by the fluctuation of the laser scanning sequence,to verify its accuracy to distinguish roads and crops,so as to verify the feasibility of the scheme.Found by the tractor navigation experiment,the tractor can accurately complete the single ridge walk and multiple ridge walking between the autonomous navigation,navigation error of the maximum is only 0.033 m higher than the accuracy of traditional navigation algorithm,in order to verify the reliability of the program.