Research On The Operation Path Recognition And Trajectory Control System Of Roller Based On Machine Vision

With the development of science and technology,the intelligent level of construction machinery represented by the roller is getting higher and higher.At present,the roller’s rolling operation is completed by manual operation,which has many problems such as high labor intensity,low operating efficiency,and the dependence of the roller quality on the driver’s driving experience.In order to realize the intellectualization and humanization of the roller’s operation,and to improve the operation efficiency and compaction quality,this paper proposes the research of the roller’s operation path recognition and track control system based on machine vision.In this paper,the analysis of the roller unmanned technology development and on the basis of machine vision technology used in engineering machinery,put forward the direction of the road on both sides of the kerbstone or other identification as a road roller operation navigation benchmark,according to the requirement of the navigation benchmark and compaction process planning out the roller operation path and uses the lateral deviation and angular deviation as navigation parameters of rolling operations.Firstly,the camera image model and the basic principle of camera calibration are used to obtain the internal and external parameter matrix of the camera through the camera calibration method of zhang zhengyou,so as to complete the camera calibration.Secondly,road images in the compaction environment were preprocessed,including ROI region setting,color image graying,image filtering,image segmentation and mathematical morphological optimization,so as to identify the operation path of the roller.Then,Canny operator is used to detect the two boundaries of the operation path,and the boundary close to the center of the vehicle body is proposed as the navigation reference line,and the modified least square method is used to carry out the straight line fitting,and a calculation model based on the lateral deviation and Angle deviation of the vehicle body in the world coordinate system is established.Finally,the feasibility of the algorithm and control strategy designed in this paper was verified through static vehicle body positioning accuracy measurement test,lens plane linear guide rail motion test and ros-based vehicle path tracking test.The results show that: under static conditions,the average measurement error of lateraldeviation and Angle deviation is 0.52 cm and 0.53° respectively.In the roller commonly used rolling speed range,navigation parameters extraction accuracy is high,and the dynamic motion of the camera is less affected;The maximum tracking error of the test car at three walking speeds of 0.36km/h,0.54km/h and 0.72km/h was 1.97 cm,2.47 cm and 2.83 cm,respectively.The tracking effect was better.The research shows that the visual navigation algorithm and control strategy designed in this paper are feasible and effective in the simulated compaction environment,and can provide a technical reference for roller and other construction machinery to realize visual navigation.