Research On Binocular Vision Work Object Measurement System For Construction Vehicles

As the most important carrying and special working tools in the engineering field,engineering vehicles have a wide range of applications.The engineering vehicle can acquire the information of the work object by means of binocular vision technology and cooperate with the work task to complete the designated work.Moreover,the size and position information of the obstacle can be obtained at the same time,and the active obstacle avoidance can be realized.In addition,because the working environment of the construction vehicle may be very bad.Therefore,the application of binocular vision technology on engineering vehicles and the improvement of equipment intelligence level are one of the main ways for unmanned engineering vehicles.Binocular vision technology is a measurement technique that is modeled after the principle that a biological eye can produce a stereoscopic feeling.It is a biomimetic technology based on image processing technology.With the rapid development of image processing technology in recent years,compared with other traditional methods,image processing technology has lower hardware requirements and can greatly reduce the development cost of measurement systems.In this paper,the binocular vision technology is used to measure the environment of the engineering vehicle,and the relative position of the engineering vehicle and the position and volume information of the obstacle can be obtained,thereby realizing the three-dimensional reconstruction thereof.The main contents of this paper are as follows:(1).Based on the analysis of the pinhole model and lens model of the current mainstream camera imaging,it is verified that the two models are approximately unified on a large scale.The imaging model for establishing the camera is then completed based on the pinhole model.And Zhang Zhengyou's calibration algorithm is used to calculate the camera's imaging model,and the experiment is designed to verify the camera calibration.(2).The left and right image brightness caused by the complicated ambient light source of the engineering vehicle and the sampling noise existing in the camera itself,and the left and right image resolution caused by the difference in depth of the left and right cameras generated by the camera focus are different.Case.In this paper,we use the histogramdistribution as the standard to equalize the brightness of the two images,and use the bilateral filtering,which can preserve the image gradient information well,to filter the image.(3).For the particularity of the working environment of engineering vehicles,that is,the stereo matching problem of small color difference and large area of low-texture area,the mutual information is constructed by image entropy value to establish the energy function to find the optimal matching point.A parallax of the pixel points is obtained to form a disparity map.(4).Using the perspective projection transformation combined with the camera imaging model to derive the re-projection matrix that projects the two-dimensional point into the threedimensional space.The re-projection matrix is then multiplied by the disparity map obtained by stereo matching to obtain the final three-dimensional coordinates.The development of the binocular visual obstacle recognition system was completed and the design experiment verified the effectiveness of the system.