基于单目视觉的农业机器人导航系统研究

Researchon Navigation System of Agricultural Robot Based on Monocular Vision

机器人在农业生产、农产品运输等方面的应用程度,已逐渐成为农业智慧化水平的体现,而受农业环境复杂性的限制,机器人的导航路径规划及定位精度问题,仍是制约农业机器人应用的主要因素之一。为此,设计了一种基于单目视觉的农业机器人导航系统,通过摄像头采集农业机器人工作环境信息,建立机器人的视觉导航地图;采用级联分类器区域检测结合颜色标定的方法,使用户能够根据具体环境,自主规划机器人运动路径,实现机器人的实时定位与导航。实验结果表明:农业机器人沿用户自主规划的无轨道路径,可自动完成导航定位工作,并在路径各目标点获得了亚米级的定位精度,满足农业机器人的应用需求。

Gradually intelligent agriculture was marked by the participation levels of robot in agricultural production and transportation. Limited by the complexity of agricultural environment,the navigation path planning and positioning resolution is still a principal consideration for applications of robots in agriculture.A navigation system of agricultural robot based on monocular vision,which collected information of the working environment through camera for building a visual navigation map of a gricultural robot. the users could the system utilized region detection of cascade classifier and color calibration,to realize the real-time positioning and navigation of robot,and the users could customize the motion path of robot independently according to specific environment.The experimental results shows it could obtain positioning resolution with sub meter level at each target point of the path,and the requirements of agricultural robot will be satisfied preferably.