Industrial Vision Systems and Defect Detection

Automated inspection systems use cameras, algorithms, and increasingly deep learning models to find flaws in manufactured materials—whether a missed weave in textiles or a pattern anomaly etched into a semiconductor wafer—faster and more consistently than human inspectors can. Getting these systems right matters because a single undetected defect in a microchip or a recurring flaw in fabric production can translate into costly recalls, wasted material, or compromised product reliability at scale. Much of the current research focuses on how to extract meaningful texture features from images—classical tools like Gabor filters remain competitive, but convolutional neural networks and other deep learning architectures are pushing detection accuracy further while reducing the need for hand-crafted feature design. Open questions center on how well models trained on one type of defect or manufacturing environment transfer to another, and how virtual metrology—estimating process quality from sensor signals rather than direct measurement—can be woven into inspection pipelines to catch problems even earlier.

Works

101,406

Total citations

724,234

Keywords

Fabric Defect DetectionMachine VisionTexture AnalysisSemiconductor ManufacturingDeep LearningWafer Map Defect Classification