Automatic Detection of Asphalt Pavement Crack Width Based on Machine Vision

Pavement disease is a key factor affecting road safety, and regular inspection and maintenance play an important role in maintaining the performance of asphalt pavement. Manual crack detection is subjective and inefficient. Even with the rapid development of pavement detection technology in recent years, it is still a challenging task to detect asphalt pavement cracks from images due to the complex environment. Crack classification and identification based on deep learning have achieved many results, but there are still deficiencies in terms of parameters and model efficiency. Aiming at the above problems, this paper proposes a new crack identification and analysis system guided by the structured information in the intelligent asphalt pavement disease image. First, the pavement crack types are classified by applying the convolutional neural network (CNN) algorithm to mine images containing 4,000 pavement breakage images and using the sparrow search algorithm (SSA) to improve the accuracy of the model. Aiming at the common problems of low contrast, serious noise pollution, and uneven illumination in the special environment of the tunnel, the improved Retinex enhancement algorithm is used to enrich the image details. Affected by the different textures of pavement materials, imaging conditions, etc., it is proposed for the first time to calculate the crack width through the skeleton boundary refinement algorithm. The results of experimental tests and case studies show that the proposed approach demonstrates the potential of innovative deep learning methods in automatically calculating pavement quality indices to optimize the repair and maintenance of asphalt pavements.