A New Method of Subway Tunnel Crack Image Compression Based on ROI and Motion Estimation

Aiming at the characteristics of the subway tunnel crack images, this paper presents a new method of subway tunnel crack image compression based on region of interest and motion estimation. It contains three key parts: the method of key frame image compression based on Discrete Cosine Transformation, the method of internal frame image compression based on forward predictive coding and motion estimation, the method of lossless image compression based on crack information database and suspected crack regions. The simulation experiment results show that this method can not only enhance the image compression ratio without losing any information of images in the region of interest, but also interface with the existing subway tunnel crack recognition system very well and make good use of the data from the crack recognition system database and the images in the disk array.