Abstract：Aiming at the problem that the pointer recognition instrument of the substation inspection robot is sus-ceptible to pointer shadow, a pointer-based instrument image segmentation method based on cubic spline interpolation is proposed. First, the circular area containing the pointer is extracted as the information map, the median filtering and grayscale stretching are performed on the information graph; secondly, the information map is segmented by using 16 fixed thresholds, and the length and area of the pointer are counted; The strip interpolation method interpolates the area sequence and calculates the first-order difference. The first threshold of the first-order difference is obtained by taking the minimum threshold of the pointer length larger than the radius of the information as the starting point, and the gray value corresponding to the point is used as the segmentation threshold. Image segmentation of the infographic. The image segmentation experiments show that the proposed algorithm not only can effectively eliminate the inter-ference of pointer shadows on instrument identification, but also extracts a small amount of invalid information, which reduces the waste of subsequent algorithms. Robust experiments show that the algorithm Environmental interference is robust.