This paper describes an image processing method to detect defects on the surface of silk screen printed plates which are plane plastics products. The final inspection of printed plates is always executed by human, and it has several problems such as oversights of defects, a long time laborious work and so on. The proposed method is developed for reduction of labor and automatic inspection. After inputting photographs of printed plates to a computer, the following processes are executed in this order; edge enhancement, position normalization, template matching and extraction of potential regions for surface defects. By analyzing features of each region, defects are extracted successfully.
View full abstract