Transfer Learning for Detection of Casting Defects Model In Scope of Industrial 4.0

Casting represents a production process where liquid material is poured into mold with hollow cavity, usually of the intended shape, following which its solidification allowed. Numerous defect types are available, including blow holes, pin burrs, defects, shrinkage metallurgical casting metal etc. All industries have quality control departments to eliminate occurrence this defective product. But main problem that inspection done manually. This very time consuming and due human sensitivity not 100% accurate. In study, we will verify whether "manual inspection" bottleneck can be eliminated by automating transfer learning in manufacturing products. images were divided two separate classes, classification was carried out applying deep architectures. The benefits proposed approach discussed as more efficient way final products under Industry 4.0.