Object Position Information Calculation Based On Two Plane Lasers And Monocular Vision

With the rapid development of related technologies in the field of computer vision,line structured light vision detection technology has attracted more and more attention.It is widely used in the actual measurement of production and life,such as three-dimensional reconstruction,intelligent manufacturing,road detection and so on.It has the advantages of low cost,high precision,good real-time performance and simple structure.It can be applied to the field of environmental perception and replace some functions of laser radar and millimeter wave radar,Achieve the purpose of reducing costs.The main research contents of this paper are as follows:(1)This paper takes the linear structured light detection system as the research object.By analyzing the working principle of the existing direct and oblique linear structured light vision detection systems,an experimental system for detecting obstacles based on two plane lasers is built;According to the actual needs,select the appropriate image sensor and line laser with performance and power consumption.(2)The calibration process of line structured light vision measurement system is studied.It mainly includes camera calibration and light plane calibration.Firstly,the four coordinate systems needed for calibration are defined.After analyzing and comparing various camera calibration methods,it is decided to use Zhang’s calibration method to calibrate the internal parameters,external parameters and distortion coefficient of the camera;Light plane calibration is to determine the spatial position relationship between the light plane projected by the line structured light and the camera,which has a great impact on the accuracy of the detection results.In this paper,the indirect calibration method is used to solve the light plane equation combined with the projection matrix obtained by camera calibration.(3)The extraction technology of laser stripe center with linear structured light is studied.Aiming at the problems existing in the existing light strip extraction algorithms,a gray center of gravity model weighted by normal neighborhood is proposed to extract the light strip center according to the studied application scenario.The experimental results show that compared with the traditional gray center of gravity method,the algorithm proposed in this paper can make up for the disadvantages of broken line of gray center of gravity method,and has higher accuracy and stability;Compared with Steger method,the proposed algorithm runs faster and has better real-time performance.(4)An experimental system based on two plane lasers is built for obstacle detection experiment.The laser stripe image is collected by the camera.The sub-pixel center of the stripe is extracted by using the normal neighborhood weighted gray gravity center model proposed in this paper to detect the distance,orientation and width information of the obstacle.The experimental results show that the above methods can accurately detect the position information of the measured object,and verify the effectiveness of the gray center of gravity model weighted by normal neighborhood to extract the light strip center algorithm proposed in this paper.