Algorithm for grading blisters of coatings: A comparison with the visual inspection

Coating appearance quality is inspected manually both in the production line and in service. Some issues regarding this visual inspection are different evaluation criteria depending on eyesight, perception, and experience of experts and various assessment criteria affected by each inspector's mood or tiredness. Since every report is affected by these human errors, Analyzing the massive data for recognizing faults in previous procedures is impossible. In addition, These problems make the description of the least acceptable quality of products impossible. Blistering is one of the common defects of coatings that occurs in the product line (mostly called bubbling) and when the paint on steel surfaces is in immersion service or exposed to high humidity for a long time. The degree of blistering of coatings is assessed based on ISO ۴۶۲۸-۲ or ASTM D۷۱۴. In both standards, ۱۶ pictures of blisters varied in size and frequency are available. Inspectors determine the blistering grade by comparing the surface with the pictures. Machine vision as a solution to the associated problems with visual inspection is used for many quality control demands. It is used in this paper to provide a fast and reliable method for assessing the degree of blistering of surfaces. For this purpose, calibration images for optical imaging systems in ISO ۴۶۲۸-۲ are used. The rate of gradation of the proposed algorithm for ۴۰ samples in four achromatic colors is more than ۹۰% by comparison with five industrial inspectors' results. Besides, It assessed each nine times faster than inspectors.

کلیدواژه ها:

Visual inspection ، Coating inspection ، ISO ۴۶۲۸-۲ ، Degree of blistering ، gradation algorithm