Research On Pavement Filling Trajectory Control Based On Machine Vision

At present,China's road traffic construction is developing rapidly.Road surface damage detection has become one of the road maintenance departments focus.In this paper,the problems existing in the process of pavement damage intelligent detection and repair trajectory control,such as the low degree of automation,the low efficiency,the labor intensity and the safty of the operator are studied.In this paper,the detection principle of pavement crack detection system is analyzed,and the modularization of the detection system is designed.The system mainly includes data acquisition module,data processing module and other function modules,including power supply system,lighting system,etc.The working principle of the whole crack detection system is to collect the pavement crack image and convert it into digital image,then use the digital image processing method to extract the cracks of the road and store and transmit the image.Digital image processing is the core content of the pavement crack detection.There are many kinds of pavement crack image processing methods.In this paper,in order to extract the trajectory of the crack,the crack image enhancment,edge detection,image segmentation,image erosion and dilation and crack refinement is processed in Matlab environment.In the process of crack image processing,different processing methods are compared to select the better processing effect.In the process of control the crack repair robot,the greedy algorithm and the optimal path planning algorithm are used to control the movement of the robotic nozzle along the crack track.Three kinds of typical pavement crack models were selected to test based on the three-axis linkage test platform with the MPC2810 motion control card.The effectiveness of the control algorithm and reliability of control accuracy were verified by simulating the trajectory of the crack repair robot.