Development of Visual Inspection for Plastics using Image Processing Technique

Abstract

This report describes the development of quality inspection of plastic products by using an ‘image processing technique’. The objective of this research is to reduce the occurrence of defect plastic parts in the production line. A visual inspection is carried out prior to the production line using a dot template film to compare black dot sizes during which there is the possibility for defect parts to leak into the production line. A Fish Bone Diagram was used to analyze the primary cause of the inspection error and a Matrix Diagram was used to calculate the relative scores between cause and effect. We found the three main causes of the problem, which are eye strength, misalignment when film is deposited on parts and light quality. Then a ‘Why Why’ Analysis was used for the analysis of the root cause of the problem and an image processing technique was applied for error observation as opposed to a film template. For making a decision easily, the defect part is divided into two types, a black dot size on a plastic part being smaller than 0.2 mm. and a black dot size on a plastic part being bigger than 0.2 mm. A comparison of ‘before’ and ‘after’ was applied to the image processing technique for inspection. It showed that we can decrease the occurrence of poor quality plastic parts in the production line from 73,017 DPMO to 1,214.79 DPMO.