To achieve rapid customization of batik patterns and improve the efficiency in innovative design of such patterns, this study studied a rapid customization system based on the related functions of the OpenCV computer vision library and the Python language. In this method, batik images were morphologically preprocessed, and the contours of batik patterns were extracted by the Canny operator. Then, the subpatterns were selected manually to encode and extract the configuration. The spatial positions of batik patterns were obtained automatically through Hu moment matching, and the configuration rules for the batik patterns were established by the adjacent graph method in topology. A reuse method of configuration rules was proposed. Patterns were embedded onto the extracted configuration rules, and the pattern embedding method based on the identity matrix method was used to avoid contour overlap. This method could enrich the design results and effectively improve design efficiency. Finally, innovative design of batik patterns was carried out through the prototype system of batik pattern design to verify the feasibility of the method.