基于复杂设施农业环境的多传感器融合建图OA

Aiming at the problem that the current 2D laser LiDAR SLAM system was not suitable for complex facility agri-cultural environment mapping and the high cost of 3D LiDAR,a mapping method based on the Ackerman agricultural ro-bot platform was proposed,which integrated 2D LiDAR,visual RGB-D camera and wheeled odometer.The multi-sensor fusion mapping model of 2D LiDAR,RGB-D camera and wheel odometer was constructed,and the SLAM mapping process of visual-radar-wheel odometer fusion slam was studied and analyzed.The proposed multi-sensor fusion mapping method was verified by experiments in simulated complex facility agricultural environment.The experimental results showed that the environment map established by this method was a fusion map of two-dimensional plane and three-dimensional space.The maximum error of fusion map was 2.2%,and the maximum error of 2D LiDAR map was 2.9%.The maximum map error of RGB-D camera-based pure visual mapping was 4.4%.The accuracy of fusion map was higher than that of 2D LiDAR map and RGB-D camera map.The maximum error of the length,width and height of obstacles in the fusion map was 16.3%,20.9%and 12.1%,respectively.And the maximum error of the distance from the centroid of obstacles to the starting point of the map construction was 4.5%,which were within a reasonable range.It met the re-quirements of automatic navigation in the complex facilities agricultural environment and effectively improved the limitation of 2D LiDAR for agricultural robot mapping in complex facilities agricultural environment.At the same time,the problems that 3D laser radar was of high cost and was not conducive to the popularization and application of agricultural robots were solved,providing theoretical basis and data support for the research of agricultural robot mapping and navigation.