A new approach for detecting and grading blistering defect of coatings using a machine vision system

Blistering is a frequent coating defect and is mostly assessed visually. In this research, a new system including an image acquisition setup in ambient lighting, a new algorithm for detecting blisters visible to normal corrected vision as well as those with a diameter of 5 mm and more, and a grading algorithm of blisters based on features extracted from ISO 4628-2 calibration images is presented. This system has efficient performance in the detection and extraction of blisters in images of achromatic colored gloss paints with both artificially and naturally made blisters. It can detect high-frequency blisters in a single shot and assess the degree of blistering of coatings with a 95 % precision compared to industrial inspectors' visual evaluation. These findings show its capability as an alternative to manual inspection for inspecting the blistering defect of coatings during production and service.