Automatic visual inspection of PCB using CAD information

A PCB is a thin board made of fiberglass with electrical wires printed onto the board, connecting the central processor to other components on the board For inspecting the PCB, the image of the bare PCB is captured and visually inspected to find the defects. In this paper we propose a system where the reference image is obtained from a CAD file. The inspection image is the captured image of the PCB which is obtained during image acquisition and is threshold. To find the defects, portion of the reference image is selected and cropped. By template matching the same portion is obtained from the inspection image. The image subtraction is performed. From the subtracted image the defects are identified. The defects are extracted by feature extraction method and by shape analysis the standard defects are colored and displayed.