Due to inefficiencies and inconsistencies in manual inspection of brake lining core defects,a machine vision-based defect detection system was developed to detect defects on the flange surface of the brake lining core.An optical scheme was designed to capture images of the lining flange surface.Parallel median filtering was used to denoise the original image,followed by the introduction of a nonlinear grayscale exponential change to make the image histogram display distinct peaks and valleys.Then,the flange area of the brake lining core was extracted by using an improved Otsu method.Finally,defect information was identified and assessed through a combination of Fourier transform and inverse Fourier transform methods.Experimental results show that the system has a detection accuracy of 98.9% and no missed detections,with an average detection time of 0.232 s per individual brake lining core part.This system offers high measurement precision and fast operation speeds,effectively replacing manual measurements in industrial inspections.