Objective To improve and optimize the geometric distortion correction algorithms for scanned images. Methods Through analysis of causes for geometric distortion in the scanner, summarization of the existing methods for image distortion correction, and comparison with the existing image processing software Photoshop, an appropriate mathematical model and algorithm flow were established. Through image preprocessing, images were enhanced by histogram equalization, edges were detected by Canny operator, and then the image contours were extracted by Radon operator, subsequently, image distortion was corrected by dot connection, and finally it was realized by Matlab simulation. Results A large number of experiments verified that this algorithm was suitable for scanned images, the algorithm processed quickly without manual processing. The similarity reached up to 0.95 after correction. Conclusion As a supplement of quick processing method for correction of scanning geometric distortion, this algorithm corrected geometric distortion of the preprocessed scanned images by dot connection, which was suitable for the geometric distortion of scanned images, with important theoretical and practical values.