Noise light spot image adaptive window preprocessing centroid algorithm and system

The invention discloses a noise light spot image adaptive window preprocessing centroid algorithm and system, and the algorithm comprises the steps: obtaining a light spot theoretical size D corresponding to a noise light spot image according to a diffraction limit angle corresponding to a measurement system; meanwhile, after the noise light spot image is binarized, a white area with the pixel value larger than a threshold value is circled, the largest frame shape is selected to serve as a standard window, the standard window size (A and B) is obtained, and A and B represent the length and the width of the standard window respectively; comparing the values of A, B and D, selecting a self-adaptive window, and obtaining the length and width of the self-adaptive window; and carrying out centroid calculation on the noise light spot image in the added window. The method has the advantages that the theoretical size D of the light spot and the actual size of the processed noise light spot image are compared, adaptiv