The work aims to research the method to detect the crosshatch-angles of cylinder liner based on improved Hough transform, in order to rapidly detect crosshatch-angles in the production process and eliminate the rejected products, with respect to the fact that the crosshatch-angle of cylinder liner is an important parameter that affects cylinder liner performance. Firstly, the image pre-processing algorithm of cylinder liner was determined, the crosshatch features of image were strengthened and the secondary information was filtered out. Then, the crosshatch feature edge segmentation method based on Canny operator and the crosshatch linear feature extraction method based on improved Hough transform were determined, respectively. The crosshatch-angle of cylinder liner was obtained by calculating the mean value of included angle between straight lines. The test showed that, compared to the detection results of imported instruments, the average error of such results of crosshatch-angle based on improved Hough transform was 1.54% and the detection accuracy was high. The proposed detection algorithm can accurately identify the crosshatch-angles of cylinder liner, properly replace the traditional manual laminating detection and the expensive imported detection instrument, meet the requirements of automatic detection on industrial site, improve the detection efficiency and accuracy, and reduce the detection costs.