Abstract: In order to solve the problem of difficult modeling due to the sparse structure of insulator point clouds and the difficulty in identifying types and anchor points, and to realize high-precision reconstruction and hooking of insulators, a method for 3-D reconstruction of insulators integrating classification, registration, and anchor optimization was adopted. Firstly, the spatial separability of insulator groups and the attitude differences between types were utilized to classify multiple types of insulators in transmission towers through clustering and fitting methods, which solves the problem that existing methods can only distinguish a single type. Then, to address the challenge of the difficulty in registering and reconstructing sparse and noisy insulator point clouds, a coarse-to-fine reconstruction method for the registration and optimization of low-quality insulator point clouds was proposed. Finally, the method of finding the nearest neighbor hanging point was employed to solve the problem of inaccurate connections between the insulator and the tower. Experiments were conducted using the above methods on different types of insulator point clouds, and the results indicate that the reconstruction accuracy of insulators can reach within 0.3000 m, with the actual shape and spatial attitude being accurately restored. The research results provide a reference for the engineering application of insulator reconstruction.