Abstract: Apparent cracks in subway tunnel linings would pose direct threats to the safe operation of subway tun? nels. Most current detection techniques for lining cracks are based on image processing, and the detection efficiency and accuracy can hardly meet practical needs due to the large amount of image data to be processed and the complex internal environment of subway tunnels. In view of the above situation, a study has been conducted on the identification method for cracks in tunnel linings based on improved chain code tracking algorithm. Firstly, the apparent images of tunnel lining, obtained by the linear array camera, were processed in advance. Then, the identified cracks were converted into chain code sequences, before an improved chain code tracking algorithm was put forward which could quickly extract the special points in the crack boundaries and eliminate pseudo-inflection points and redundant inflection points. Lastly, based on the tree structure model of crack skeleton, algorithms for crack skeleton length and crack width, were proposed and an error analysis was carried out for the chain code tracking results. The results of field detection experiments show that the number of pixel points of the crack boundaries tracked by the improved tracking algorithm of chain code boundaries can be reduced by 82%, with the errors of crack length and width controlled within 4% and 7% respectively, under the premise that the crack shape remains unchanged. Therefore, this method can provide technical support for the rapid detection of cracks in tunnel linings and the safe operation of tunnels.