Learning Vector Quantization Neural Networks for LED Wafer Defect Inspection

Automatic visual inspection of defects plays an important role in industrial manufacturing with the benefits of low-cost and high accuracy. In light-emitting diode (LED) manufacturing, each die on the LED wafer must be inspected to determine whether it has defects or not. Therefore, detection of defective regions is a significant issue to discuss. In this paper, a new approach for inspection of LED wafer defects using the learning vector quantization (LVQ) neural network is presented. In the wafer image, each die image and the region of interest (ROI) in them to handle can be acquired. Then, by analyzing the properties of every ROI, we can extract specific geometric features and texture features. Using these features, the LVQ neural network is presented to classify these dies as either acceptable or not. The experimental results confirm the usefulness of the approach for LED wafer defect inspection.