Abstract: In order to improve the quality of coil winding and detect the winding defects such as gap errors and overlap errors in high-precision winding operations, a method was proposed for real-time detection and modification of different winding states in the winding process. Considering the winding characteristics of the coil, a real-time detection algorithm was proposed for winding defects of the enameled wire, and a corresponding visual detection system and control system were designed for the winding equipment. Based on the 1×3 template, an extraction algorithm was developed for the contour point extraction of the preprocessed effect images. Based on the winding state detection of the contour point fitting, a modification method was provided to realize the real-time detection of the coil winding state, and to modify the wrong winding state according to the results of visual real-time detection. The verification results of visual detection stability test, system accuracy rate test and orthogonal test show that the proposed method can obviously increase the tightness of the coil winding, and effectively improve the production efficiency and quality of the coil winding products.