This paper analyzes the image processing scheme of the capture detector in a spot position detection system based on a four-quadrant detector. Capture unit image by using statistical averaging method for noise template production, combined with the differential shadow method to remove the background noise. The spot center of mass coordinates are obtained to capture the beam. The optimized median filtering scheme, twodimensional Otsu algorithm are used to process the laser spot image respectively, and then Canny operator is applied to extract the edge of the spot, which achieves the detection and localization of the target spot, and prepares for the initialization of the spot tracking algorithm. The related filtered light spot tracking algorithm is proposed, and the operation process is arranged for it. The spot position detection test conditions are built, and the related filtered spot tracking algorithm is experimentally verified and analyzed. The root-mean-square difference of the correlation filtered spot tracking algorithm for linear and curved trajectories is less than 5 mm, which shows that the correlation filtered spot tracking algorithm expands the detection range of spot position and improves the detection accuracy of spot position without increasing the complexity of the algorithm.
