Target Recognition For Autonomous Following Of Mobile Robots

In this paper,according to the application requirements of autonomous tracking of ground mobile robot,the method of target recognition and location based on multi line laser radar is researched.This paper introduces the research status of ground mobile robot technology and presents a method of target recognition.First,collecting 32 line laser radar point cloud data,and convert data from polar coordinates to Cartesian coordinate system,building grid map put 3D data onto a two-dimensional horizontal plane,using four neighborhood algorithm and adjacent point clustering algorithm to cluster the grid map,establishing the box model to express the obstacles of the same class.Achieving detection and classification of the obstacle.Then,using the difference method to detect the moving target,and according to the geometric characteristics of pedestrians and vehicles remove excess noise,achieving initial location of the target;the target area is projected onto the vertical plane,using the human geometry model to achieve recognition of pedestrian target;at the same time,a matching method is proposed,which includes the body reflection intensity and the license plate reflection intensity and the zero reflection intensity to achieve recognition of vehicle target;finally,calculate the location information of target.Finally,in order to verify the proposed target recognition and location method,constructing the experimental system of data acquisition and processing for multi line lidar,write the target recognition and location procedures to conduct the verification experiment of indoor and outdoor environment,single person,multi people,pedestrians and vehicle and other scenes,the experimental results show that this method can accurately identify the location of people and vehicles.