This paper analyzes the disassembly and installation of busbar fittings bolts for substation robots in 110 kv environment, and proposes image self-correction, template segmentation fitting arc and panel pose correction for visual positioning of busbar fittings. The image self-correction solves the quality problem of the fitting image acquisition of the complex illumination condition by rectifying the overall of dark images and the local of bright images. The template segmentation fitting arc method performs multiple segmentation and arc fitting techniques on the fitting template image to accurately calculate the centroid of the fitting, thereby completing the visual positioning of installing bolt. The panel pose is solved by disparity map. Through the panel pose correction, the binocular camera can take photo vertically and use template segmentation to locate the bolt, thereby completing the visual positioning of removing bolt. The appeal method is based on the specific conditions of the disassembly and assembly of the busbar fittings, which solves the problem of large changes in ambient illumination during the visual positioning of the busbar fittings and the uncertainty of the shooting pose.