高反光场景下高压电缆接头表面缺陷的偏振视觉检测OA

To address the issue of low visual inspection accuracy caused by high specular reflection on the surface of high-voltage cable joints,this paper proposes a visual detection method that integrates polarization imaging with defect-specific differentiation.The method extracts the degree of linear polarization(DoLP)and angle of linear polarization(AoLP)from Stokes vectors and fuses them with weighted combination to suppress high-reflection interference and enhance defect separability.Aiming at two typical surface defects-residual semiconductive particles on the main insulation and incomplete peeling of the outer semiconductive layer-a defect-specific differentiation algorithm is developed.This algorithm combines adaptive threshold segmentation with morphological optimization for one defect type,and sub-pixel edge extraction with random sample consensus(RANSAC)-based conic fitting for the other,enabling localization,segmentation,and type classification of the two defects.Experimental results demonstrate that the proposed method achieves detection accuracies of 95.7％and 94.3％for the two defect types,respectively,with both false-negative and false-positive rates lower than those of conventional visible-light imaging methods.This work provides technical support for quality inspection of high-voltage cable joint fabrication.