سال انتشار: ۱۳۹۰
محل انتشار: هفتمین کنفرانس ماشین بینایی و پردازش تصویر
تعداد صفحات: ۶
Parisa Kamani – Amirkabir university of Technology
Elaheh Noursadeghi – Amirkabir university of Technology
Ahmad Afshar – Amirkabir university of Technology
Farzad Towhidkhah – Amirkabir university of Technology
Nowadays on line automatic inspection plays animportant role in industrial quality management. This paperproposes a new computer vision system for automatic paintedcar body inspection in the context of quality control inindustrial manufacturing. In most worldwide automotiveindustries, the inspection process is still mainly performed byhuman vision, and thus, is insufficient and costly. Therefore,automatic paint defect inspection is required to reduce thecost and time waste caused by defects. This new systemanalyzes the images sequentially acquired from car body todetect different kinds of defects. Initially, defects are detectedand localized by using a joint distribution of local binarypattern (LBP) and rotation invariant measure of the localvariance (VAR) operators and next, detected defects areclassified into different defect types by using Bayesianclassifier. The results show that this method could detectdefects and classify them with high accuracy. Because of itssimplicity, online implementation is possible as well.