| Summary: | 碩士 === 南臺科技大學 === 電子工程系 === 106 === Electronic products are more and more complex and precise nowadays, so their printed circuit boards (PCBs) must be inspected before they’re shipped out. The Automated Optical Inspection (AOI) is a reliable technology for automatic PCB inspection. This thesis aims at designing an AOI prototype system for PCB inspection. The hardware is composed of three moving stages adapted from a 3D printer, RGB light sources, a CMOS image sensor. The software consists of two parts: the image inspection of PCB components and the user interface. The major inspection items are part appearance, part number, part location, solder joints, pins of integrated circuit. The user interface can be used to set location, item, parameter, and report errors. The user interface will show inspection results, so the user can set correct parameters.
The RGB light sources irradiate at different angles. The CMOS camera’s resolution is three million pixels and is used to take images of component surface and solder joints. The images will be preprocessed by Gaussian blurring, contrast/brightness adjustment etc. The preprocessing is used to increase inspection accuracy and transform image into appropriate format such as image binarization, Hue, Saturation and Lightness(HSL). The preprocessed images will be compared with a set of standard images using methods like optical character recognition(OCR), template matching and color discrimination. The comparison results will be shown on the user interface of AOI software after inspection is finished.
Our AOI system uses an effect and accurate algorithm to inspect PCB parts and solder joints. The system hardware is composed of low-cost easy-to-assemble components. The system software has a friendly human–machine interface to set parameters and run inspections. The software is made up of modular functions to take advantage of modifying and adding functions easily in the future.
|
|---|
