| As the development of Surface Mounting Technology towards to lightly, thin and micro-miniature, traditional inspection arithmetic can not satisfy productive demand so that Automatic Optical Inspection (AOI) has gotten an extensive investigation. AOI means that digital image of inspecting object is obtained by optical imaging technology, and then the image is calculated by computer for inspection, analysis and determination the object.Aiming at principal inspection object, it's designed collectivity structure of system. In the hardware of AOI, the design of illumination system adopts lamp-house, annular fiber laser and radial intensity auto-regulation. Charge coupled device of planar matrix is chosen for video camera and the image collection card is also chosen to correspond to the camera.In this inspection algorithm, firstly, the PCB image is processed for noise canceling; When taking the image of PCB, the x,y transporter is not always precise so that the image may be deflective or rotational, Prewitt arithmetic operators is used for detecting border. Then coordinate system of the image is located again by searching strategy in three directions and Hough conversion. This subject presents a region-growing algorithm based on the initial point selection in windows and adaptive growing threshold, to segment the element area from PCB picture automatically. The selection of the initial point in region-growing algorithm is improved. And the growing thresholds are determinated automatically by an approach which grounds on statistical data of the area near the initial point and the feature of the grown region. The partial feature and global feature are considered in the process of element segmentation. Finally, the border of element is by least-squares algorithm, and then the task of surface mounting inspection is finished.AOI system is compiled by Visual C ++ in windows operating system. The result shows that AOI system has big practical value because of high speed and wild inspecting range. |
PDF Full Text Request