摘要
为适应农业机器人在半结构化、半开放式农业场景的复杂环境区域下,GPS导航可能信号缺失、精度偏低,导致导航定位和建图的精度偏低,不能准确避障,甚至对农作物和人员造成损伤。笔者针对单一传感器在农业机器人目标检测及机器人建图中的局限性,提出了一种利用16线激光雷达和相机融合的目标检测算法,设计了一种多传感器融合的目标检测及基于elevation mapping的三维建图。仿真结果表明:LiDAR和摄像头传感器数据可以提供深度和颜色信息,并通过elevation mapping算法实现建图,实现了农业机器人识别和无碰撞导航;使用多个传感器来提供冗余信息,以减少发生错误测量的可能性,解决了农业机器人在复杂环境区域中行走及定位检测的问题。
In order to adapt to the complex environment of agricultural robots in semi-structured and semi-open agricultural scenarios,GPS navigation may lack signals and have low accuracy,resulting in low accuracy of subsequent navigation,positioning and mapping,inability to accurately avoid obstacles,or even damage to crops and people.Aiming at the limitation of single sensor in target detection and robot mapping of agricultural robots,the author proposes a target detection algorithm using 16-line lidar and camera fusion,and designs a multi-sensor fusion target detection and 3D mapping based on elevation mapping.The simulation results show that LiDAR and camera sensor data can provide depth and color information,and map building is realized through elevation mapping algorithm,and agricultural robot recognition and collision free navigation are realized;Multiple sensors are used to provide redundant information to reduce the possibility of error measurement and solve the problem of walking and positioning detection of agricultural robots in complex environmental areas.
作者
肖正邦
Xiao Zhengbang(School of Engineering Machinery Chang’an University,Shaanxi Xi’an 710064)
