In order to solve the problem of robot positioning of bolts on transmission lines, an algorithm for detecting the edge and pose of bolts on transmission lines based on multivariate fusion is proposed. The bolt image is preprocessed, the edge is detected by Canny module, the noise is removed by connected domain analysis, and the complete bolt contour is extracted by Graham scanning module. Through a large number of experiments, the algorithm is compared with other detection algorithms, and the results show that compared with the other four algorithms, the algorithm can better detect the bolt contour in the transmission line equipment, and can still maintain a good advantage when the bolt background contrast is low, and can locate the manipulator according to its pose detection algorithm.