Automatic recognition for catenary insulators of high-speed railway based on contourlet transform and Chan–Vese model

Aiming at the diversity distribution of the catenary rod insulator image, this paper proposes a new idea of detecting and recognizing the rod insulator of high-speed railway, which is based on contourlet transform (CT) along with Chan–Vese (CV) model. First, the anisotropy and directionality of the catenary rod insulator image is detected with CT in the frequency domain, which can provide a variety of texture analysis. Then, in order to find the center of textures, the digital morphology filtering and the fuzzy clustering are adopted to process the textures. Finally, the insulator boundaries, including the abundant geometry information, are detected by CV model. The experimental results show that the proposed approach can efficiently realize the automatic recognition of the rod insulators in real-life catenary images of high-speed railway, even with the insulator porcelain skirt interdigitating.