Sand-bed defect recognition for 3D sand printing based on deep residual network

The 3D sand printing (3DSP), by binder jetting technology for rapid casting, has a pivotal role in promoting the development of traditional casting industry as result producing high-quality and economical molds. This work presents an approach monitoring analyzing powder sand-bed images to serve realtime control system 3DSP machine. A deep residual network (ResNet) is used classify defects occurring during spreading stage process. Firstly, pre-trained was applied initial parameter; then it fine-tuned on labelled defective sample dataset accomplish task, which defines induced processing. Furthermore, recognition positioning were readily achieved dividing into blocks. Experiments show that 98.7% classification accuracy validation 95.4% images.