Research On Image Processing Of In-vehicle Tunnel Lining Crack Detection

The high-speed rail tunnel is a complex narrow space used for rapidly carrying. The limited space decide the difficulty and complexity of the corresponding management measures.Traditional lining crack detection methods are artificial visual inspection and manual detection equipment.Owing to the density of high-speed railway and the limited maintenance time, the traditional methods could not complete the inspection task within the stipulated time. It is necessary to develop a rapid and efficient automated detection system.First, basing on a full analysis of the road and tunnel information collection system at home and abroad, the principle applied for the in-vehicle high-speed rail tunnel lining cracks automatic monitoring system is introduced. Then,the tunnel image acquisition system which consists of multiple CCD cameras is designed preliminary.Second, the analysis of image processing of the pictures collected by the automated detection system, as well as image smoothing, edge detection. Various filtering algorithm are compared with edge detection operator, then choosing the most suitable image processing methods.Again, the characteristic values of images which processed completely are read, and the cracks will be chosen as the representative of the characteristic value. Due to the diversity of cracks, a reading method of the cracks of image is proposed.Finally, summed up the error of the automatic image detection, as well as inadequacies of the system and the image processing, which will provide a guideline for subsequent improvement and research.This paper comprehensively introduces the composition of the automatic image detection for rail tunnel lining cracks and issues related to subsequent image processing, providing certain basis for high-speed tunnel detection system developed and efficient information extraction.