A Hybrid Defect Detection Method for Wafer level Chip Scale Package Images

In the majority of semiconductor manufacturing, the visual inspection process of the wafer surface depends on human experts. However, the inefficiencies of human visual inspection has led to the development of image process to perform inspection tasks. The occurrence of different type of defect arises from the manufacturing processed variations, like miss-out calibration or poor maintenance of the equipment. There are several semiconductor inspection approaches have been proposed but the performance is limited by the variations of defect which has different surface shape, texture, intensity, size, etc. Automated inspection methods that have been developed are tuned for the specific object and background in image processing methods. In our case, the problem of defect detection is even more complex. There are different kinds of texture such like, complex surface, variation defect and bright spots which are metal reflect light, the weak spots which are caused by dust and low intensity defect, in our test image. In this paper, we propose a hybrid inspection approach to detect defect in chip. For high intensity defect, a simple threshold method is firstly used to detect candidate defect which contains defect, bright spots, weak bright spots and noise by the characterization of high intensity. The morphology operations are then used to filter out small candidate defect which is noise. Finally, the support vector machine is used to classified high intensity defect. For low intensity defect, a boundary defect detection algorithm is presented to detect low intensity defect. Experimental results demonstrate the effectiveness and efficiency of the proposed method.