X-ray Image Enhancement Segmentation And Measurement Techniques For Components Packaging Process

Integrated circuit industry is important to the economy of a country or region. With thedevelopment of IC(Integrated Circuit)’s integration and packaging techniques, the needs forinternal defect inspecting of packaging components are increasing. The bubbles in thepackaging components seriously affect the electrical properties and mechanical properties ofthe packaging components. X-ray image can help to detect the IC’ internal defects whichtraditional Automatic Optic Inspection(AOI) can’t detect. To solve the problem of X-rayimages’ low contrast, strong noise and arbitrary shape, the following works are studied in thispaper:1. By analyzing X-ray images’ capturing principles in IC production process, the reasonsof low contrast and strong noise of IC’ X-ray images are obtained.2. To suppress the strong noise of IC’ X-ray images, Gaussian filter and image denoisingmethod based on compression perception are introduced. Experiments showed that Gaussianfilter can filter the noise of IC’ X-ray image better than another method. But both the methodswould make the edge of the X-ray image dim.3. Image enhancement methods based on the fuzzy set and the LAHE (Local AdaptiveHistogram Equalization) are proposed, in which the method based on the fuzzy set canenhance the image contrast without changing the ranking relation between adjoining pixels.4. Bubble detecting methods with and without the template based on Canny andmathematical morphology are employed to obtain the edges, the information of area andcentroid of bubbles.Experiments showed that the combination methods by Gaussian filter, enhancing methodbased on the fuzzy set and Canny algorithm can be better to realize the detection of theintegrated circuits’ internal defects and reach the testing requirements of the production.