A lower envelope Weber contrast detection algorithm for steel bar surface pit defects

A lower envelope Weber contrast (LEWC) recognition algorithm is proposed to detect steel bar surface pit defects. Some present algorithms could detect these defects but they are limited by the size of pits. This kind of defect has uniform background intensity in column pixels of steel bar surface images while it is not the case in row pixels. The lower envelope of column pixels is used to eliminate the effect of high gray level peak value points, and Weber contrast is employed to make that all areas in one image have the same threshold to detect pit defects. Experimental results show that the proposed LEWC algorithm has a high detection accuracy rate which exceeds 98% for the pit defects with different sizes.

► A lower envelope Weber contrast (LEWC) detection algorithm for steel bar surface pit defects.
► Weber's Law is applied to computer vision defect detection.
► The Weber contrast of column pixels' lower envelope signal is calculated.
► The LEWC algorithm has a high detection accuracy rate.