Research And Implementation Of Pavement Crack Path Planning Method Aided By Machine Vision

In the past 40 years,China’s road transport industry has developed greatly,with the total highway mileage reaching 4.9595 million kilometers.The increase of the total highway mileage poses a challenge to highway maintenance,the sealing of pavement cracks,the repair and excavation of sewage pipes all require advanced,efficient and highly automated highway maintenance equipment and machinery.When developing such remote-operated and machine-assisted systems,path planning is the primary task of robot motion,regardless of whether their environment is structured or unstructured.In this paper,a path planning algorithm and a single crack sealing method for automatic pavement crack filling system with machine vision are proposed to provide theoretical support for realizing a more automatic,portable and intelligent automatic pavement crack filling machine in the future.This paper mainly studies the following aspects:1)In this paper,the performance of the existing greedy path planning algorithm and the two-phase tree path planning algorithm are compared,in which the greedy path planning algorithm has been applied to the existing road filling system.2)Based on the greedy algorithm and two-phase tree algorithm,a new path planning algorithm--an improved greedy path planning algorithm based on road cracks is proposed.In terms of reducing the computation time of the algorithm,the problem that the more the number of cracks,the longer the time of the two-phase tree algorithm is solved is solved.In the aspect of reducing the idle path error,the problem that the greedy algorithm has a bigger gap with the increase of the number of cracks is solved.Finally,the calculation time and idle distance of the three algorithms are compared,and the conclusion is drawn: by setting the crack threshold,the three algorithms are respectively used for idle path planning,and the results show that the work efficiency of the automatic pavement crack filling system is significantly improved.3)After finished the free path planning,the movement path of pavement crack sealed with pouring gun is studied.And then based on the traversal of a single crack DFS and BFS algorithm,cracks in the self-built crack-Data aspirant traveled through the operation of the basis of the Data set,the penetrance,effective branch scores and computing time as evaluation parameters of algorithm performance,the final test results show that BFS algorithm is more accurate than DFS algorithm.After that,we combine the idle path algorithm and the breadth first search algorithm into a whole model,the running time of the whole model and the total distance of the slot gun are obtained.However,the accuracy of the existing pouring guns is not enough to meet the requirements,so the method is suitable for the future when the precision of the pouring guns reaches the level of pixels.