Research On Detection Method Of Abnormal Areas Of High-density Board

Because of the high-density circuit module with high density, high reliabilityfeatures, and a series problems of difficult assembling, low rate of qualified products,high failure rate, there is a great challenge for traditional testing processing and methods.Now for a lack of effective means of detection, the detection of high-density circuitmodule is difficult and inefficient, while single testing means are inefficient andtime-consuming to ensure the accurate position of the fault. So scholars are taking moreconcern and making more research on the new detection methods.The paper analyzes the key technologies of the electromagnetic scanning detectionand study the judgment of the abnormal areas in detail. The outline of abnormal areas isgot by the hypercomplex space of the color image edge detection. Firstly, the Prewittoperator is improved via the genetic algorithm, and then the results are used in thehypercomplex space of the color image edge detection in order to suppress noise andincrease the accuracy of the results. The effective areas of the threshold image areextracted by the algorithm, and the areas within the outline are the abnormal areas thatwe need. Secondly, the image fusion algorithm of Gaussian-Laplacian pyramiddecomposition is improved based on the features of the threshold image, and thethreshold image and the design document image are integrated with one image，whichlays a good foundation for the subsequent determination of the suspected faultycomponents. Finally, the suspected faulty components are determined ultimately.Theoretical analysis and experiment results show that it is feasible and effective tointroduce electromagnetic scanning technology into detection of the high-density circuitboard. Moreover, the algorithm of the image edge detection which is proposed in thispaper is both accurate and efficient, and it also retains more effective information thanthe common gray image edge detection and the color image edge detection ofhypercomplex; the synthetic image that is got by the algorithm of image fusion not onlyretains the color features of the threshold image, but also the component information inthe abnormal areas is seen clearly form the synthetic image. It lays a good foundationfor the subsequent determination of the suspected faulty components.In the following research work, it is necessary to optimize testing process of thewhole high-density circuit. Meanwhile, the algorithm of determination of suspectedfailure components should be researched.