Design the Abnormal Object Detection System Using Template Matching and Subtract Background Algorithm

Computer vision is an artificial intelligence (AI) subfield that enables computers and systems to extract information from digital photos, movies, and other visual inputs. Detection systems are an important use in industrial production lines. In this paper, an automatic small abnormal object system is designed. First, the author obtains an image devoid of anomalous objects, which is then processed using the Candy filter to produce the standard form. Second, define the primary pattern, sub pattern 1, sub pattern 2, and the deviation between the new image and the original image. Then, we use template matching and background subtraction to identify questionable locations. Finally, live picture features will be compared to original image features. With Candy filter, the precision will be enhanced. The findings of image processing will be transmitted to operate the automatic abnormal removal equipment. The result indicates an accuracy of ~ 90%. The processing time is < 5 s, which has no effect on the production line cycle time.