Research On Multi-target Localization Method Based On Planar Target

The development of science and technology drives the continuous improvement of robot technology.At present,the application of multi-robot systems is gradually becoming popular and practical.Accurate positioning of the robot is to ensure that it can work efficiently,which is of great significance for improving the efficiency of the robot and promoting the level of modernization.Machine vision measurement technology is an emerging measurement technology with significant advantages such as non-contact,instant measurement,and rich information.It has been widely used in industrial production and robot positioning.Aiming at the problem that the current multi-robot system is difficult to achieve multi-target positioning,this thesis is proposed a method for positioning multiple robots using machine vision measurement technology.Firstly,a multi-target positioning system is established,which is composed of a computer,a camera,a two-dimensional turntable,wireless communication devices,and multiple planar targets.The plane target is installed on the robot to for positioning.The system uses wireless devices to light up different plane targets,and then uses vision measurement technology to calculate the pose of the plane target.The camera uses a small field of view for imaging,and is installed on a two-dimensional turntable to expand the field of view,which can achieve high-precision positioning of multiple targets in a large range.This method has theoretical significance and practical application value.The research work completed in this thesis is as follows:(1)The multi-target positioning system is constructed and its working principle is studied,the coordinate transfer law of the imaging target is analyzed,the imaging model of the multi-target positioning system is established and the pose parameters of the plane target are calculated.(2)The system communication was designed,including the structural design of the plane target and the software flow design between the coordinator and the terminal node in the communication network.The communication test of the system was carried out,and the test results verified that the system communication was in good condition,and effective data transmission and reception could be completed between each node.(3)The calibration of the multi-target positioning system has been completed.The method for detecting target points on the target is designed,which involves brightness screening,image denoising,threshold segmentation,morphological operation and feature point judgment.(4)The program has been written to complete the multi-target positioning system software.The principle experiment proved that the method in this thesis is feasible,and then the positioning accuracy of the system was compared.The experimental results show that the target positioning accuracy of the camera combined with the two-dimensional turntable and the camera without the camera are basically the same.The system also effectively expands the positioning range and can achieve more target setting.Finally,the reasons for the errors in the positioning accuracy of the system are listed.