Scratch enhancement method based on Gaussian probability density function

The invention belongs to the technical field of image processing, and particularly relates to a scratch enhancement method based on a Gaussian probability density function, which comprises the following steps of: obtaining position information of a scratch through a projection method, and marking an approximate spatial position of a scratch area according to the position information; the scratch area is subjected to block processing, the small block area is set to be square, the side length is the width of the scratch, the strip-shaped scratch is decomposed into a plurality of square detection areas, and each square area possibly comprises a section of scratch; performing scale transformation on each square area by using a weight function model to enhance scratches; the mean value T of the enhanced square area is calculated, the offset C is set, if the offset C is smaller than the threshold value, the mask is scratched, the area is traversed to obtain the scratched mask, and by means of the method, the enhanc