Research On Algorithm Of Alphalt Pavement Crack Detection And Classification Based On Spatial Features

Crack is the most important factor affecting the performance of road surface,and it is the core of daily road maintenance work.Currently,visualization technology is used at home and abroad to achieve fast and continuous acquisition of high-resolution grayscale image data on the road surface.However,the processing of pavement image crack data can only be based on man-machine interaction.It has low efficiency and poor reliability,and it has become a restrictive highway.Road network-level rapid maintenance of technical bottlenecks.In view of the problems of low efficiency,poor accuracy,and limited adaptability of automatic treatment of diseases at home and abroad,this paper studies the identification and classification methods based on the spatial distribution of cracks,so as to achieve high reliability of cracks in asphalt pavements based on gray data.High-precision and high-efficiency automatic recognition and classification.The main research content is as follows:(1)By analyzing the spatial distribution of cracks in two-dimensional grayscale images,the global discrete and local distributed aggregation strategy of crack data is proposed,and the process of pavement crack identification and classification algorithm based on spatial features is designed.(2)A method for extracting the confidence region based on sub-block partition is proposed.The adaptive block method of road surface gray image is studied to eliminate the influence of uneven brightness of the image;based on the uniform segmentation and non-uniform segmentation of sub-blocks Based on block dynamics,the rough segmentation threshold is calculated to obtain crack-like data,which provides a reliable basis for subsequent processing.(3)Sub-block fine segmentation of the crack confidence region is performed using the neural network method,and the fracture sub-section is obtained by combining the spatial distribution characteristics.Further,an evaluation model is established based on the confidence score of the sub-segment of the maximum comprehensive weight,and a method for evaluating the confidence of the crack sub-segment is proposed to realize the reliability identification of the complete crack.(4)Based on the distribution characteristics of fracture space relationships,a fracture type classification algorithm is designed based on feature mapping and principal component analysis cracking.On the basis of strengthening the characteristics of fracture types,redundant information is removed and the spatial characteristics of fracture types are extracted.At the same time,using the clustering algorithm to achieve the initial classification of crack types through the extracted features,the accuracy of crack classification is improved.In this paper,the algorithm is verified in nearly 100 kilometers of asphalt road detection,and the effectiveness of the asphalt crack recognition and classification algorithm based on spatial features is verified.Experiments show that the algorithm has strong environmental adaptability.There be a good application prospect and practical value.