Road Crack Enhancement Method Based on Hessian Structure Analysis

Abstract

The invention provides a road crack enhancement method based on Hessian structure analysis, which includes preprocessing the road surface image to obtain the original image to be enhanced for road cracks; analyzing the Hessian structure features, including obtaining the Hessian structure of each pixel on the original image ; Carry out feature analysis on different structural targets of the original image, including obtaining the eigenvalue and eigenvector of the Hessian structure corresponding to each pixel on the original image; crack enhancement, including enhancing the brightness of the crack target on the original image, weakening the background target and The brightness of the pseudo-crack target is obtained to obtain an enhanced image; the enhanced image is subjected to morphological analysis to eliminate burrs and isolated noise points on the edge of the crack, and the final enhanced image is output.

Description

technical field [0001] The method belongs to the technical field of image processing and pattern recognition, and in particular relates to a new multi-scale road crack enhancement method based on Hessian structure analysis, which can significantly enhance the contrast near the crack area on the road image. Background technique [0002] During the use of the road, it is inevitable that some large and small cracks will appear. These cracks not only reduce the service life of the road, but also pose a threat to traffic safety. Therefore, road maintenance is an urgent task. Pavement crack detection is the primary condition in the pavement maintenance process. According to the traditional manual detection, it is not only time-consuming, but also manpower and financial resources. With the development of automation, intelligence, and high-speed technology, how to quickly and accurately detect and identify cracks in road crack images has become a hot issue in the field of digital im...

AI Technical Summary

Problems solved by technology

So far, there have been some studies on fracture enhancement at home and abroad, but most of them use a single scale to apply nonlinear finite differential operators in a series of directions; Orkisz proposed a method of median filtering along the fracture direction, but this The method only operates on a single scale and cannot adapt to large scale changes

Tang Guowei et al. used region growing technology to segment road images. Although it has some effect, the amount of calculation is too large and cannot be well applied to the actual system.

Images