Abstract: The damage to the oxide layer caused by laser cleaning is visually very subtle compared with the unda-maged oxide layer,and the contrast between the damaged area and the background is low, making it difficult to distinguish clearly in the original image,resulting in a decrease in the accuracy of damage recognition. A new method was proposed to solve this problem, and the influence of laser cleaning effect on the identification of surface oxide layer damage on stainless steel parts was analyzed. The Otsu algorithm was used to automatically calculate the optimal global threshold,determine the image of the damaged area of the oxide layer on the surface of stainless steel parts,preliminarily divide the image into oxide layer and non-oxide layer,and separate the oxide layer from the overall image. The watershed algorithm was employed to address the boundary ambiguity issue in the initial Otsu segmentation,enabling precise segmentation of oxide layer regions that are in contact or overlap with each other,while eliminating interference from non-target areas. Differential operation was performed between the original image of the oxide layer on the surface of stainless steel parts and the image after nonlinear diffusion. Based on the standard deviation of image blocks in the differential image,an adaptive binarization model was constructed to perform binarization processing on the differential image. While preserving important edge information, background non-uniformity was suppressed,thereby achieving damage identification of the oxide layer on the surface of stainless steel parts and effectively improving the recognition performance. Experimental results demonstrate that the proposed method can accurately identify surface oxide layer damage on stainless steel parts.