Title :
Improvement of auto panel defect detection algorithm
Author :
Zhou, Chunqiao ; Jin, Xuanhong ; Dai, Shuguang
Author_Institution :
Coll. of Opt.-Electr. & Comput.-Eng., Univ. of Shanghai for Sci. & Technol., Shanghai, China
Abstract :
According to the problems that there are false and missed detections during the detection of non-uniform brightness of the image based on template matching algorithm, this article proposes a method that combines the two algorithms of template matching and calculating the pixel areas of the connected domains to find the image defect regions. This new algorithm´s first step is to split some regions of interesting (ROI) from the whole testing image, and we can calculate the average pixel areas of the ROI by using the image algorithms of feature extraction and regional analysis. Based on the the average pixel areas comparison with testing regions´ pixel areas, we can easily judge if the testing images have defect parts. Experimental results show that the method has a high testing precision and a quick regions positioning speed, besides, it also has a very good market value.
Keywords :
automobile manufacture; brightness; feature extraction; image matching; inspection; production engineering computing; ROI; auto panel defect detection algorithm; automobile manufacturing technology; average pixel area calculation; average pixel areas comparison; feature extraction; image defect region; nonuniform brightness detection; regional analysis; regions of interesting; template matching algorithm; Algorithm design and analysis; Correlation; Detection algorithms; Filtering algorithms; Morphology; Noise; Testing; connected domain; feature extraction; regional analysis; regions of interest; template matching;
Conference_Titel :
Fuzzy Systems and Knowledge Discovery (FSKD), 2012 9th International Conference on
Conference_Location :
Sichuan
Print_ISBN :
978-1-4673-0025-4
DOI :
10.1109/FSKD.2012.6234263
