A robust methodology for solder joints extraction

In Electronic Manufacturing Industry, machine vision systems have been announced to outperform the electrical inspection systems effectively. It supports the Surface Mount Technology (SMT) and improves the diagnostic capabilities. The challenge there is to miniaturize components with high packing density under economic considerations. This paper presents a front-end automatic detection system tackles with the solder joint specularity, illumination variations and recognition misalignment problems. This can be achieved by enhancing the threshold-based segmentation method using Discrete Cosine (DCT).