Research On Location And Recognition Technology Of PCB Components

With the progress of electronic equipments’technology of design and producting, printed circuit board (PCB) also becomes high-density, multi-layers and high-performance and so on. The detection of traditional PCB components is mainly by artificial visual inspection or function testing to achieve, but such methods have been unable to meet the high speed and precision testing requirements on production line. Because of this, this paper is the research about the detection of the chips on PCB, to design an intelligent software basing on MATLAB about PCB location and identification, by using image processing technology, completing the locating of the chips and identifying the characters on the chips, to reach the actual production of the high requirements for PCB chips detection, while reducing costs.PCB image pre-processing expounds detailedly in this paper, including image enhancement、grayscale change and image correction. Use image filter to remove the random noise; By observing the histogram, the chips and PCB’s peaks of grayscale are close, not conducive to binarization, so change the grayscale of them, to make the two goals’peaks separated; The PCB image may be a little inclined when the camera takes a picture, so must correct the image by using the method of Mark point correcting, and use the Mark point as the origin of coordinates, to establish a Cartesian coordinate system, is good to chips of the subsequent locating and identification. Taking erosion and dilation processing, remove small target area and noise, and making the outline of objects rounded and smooth.For identifying chips, use the algorithm that preliminary location according to the targets’edge pixel mutation, then precise location by four-neighbor areas searching to determine the location, then record the center coordinates of the chips, save as a standard library or compare with the standard library. For characters recognition on the chips, first of all, initialize the templates, the initialization includes classifying and coding, this improves the efficiency of retrieval, then the testing characters do the same thing, according to the classification, compare with the codes of the same templates, achieve the final result, save as a standard library or compare with the standard library.This paper uses the commercial image processing software MATLAB producted by The Math Works of US company to develop the PCB intelligent chips locating and recognition software. This software is simple, high accuracy and friendly interface.