LPViT: A Transformer Based Model for PCB Image Classification and Defect Detection

PCB (printed circuit board) is an extremely important component of all electronic products, which has greatly facilitated human life. Meanwhile, tons PCBs in the waste streams become a resources, puts recycling and reuse urgent need. In manufacturing classification PCBs, recognition sub-components, defect detection have been key technology. Traditional manual are subjective rely on individuals’ experience. With development artificial intelligence, lots research efforts dedicated to automated PCBs. this paper, we propose transformer-based model, LPViT, for We conduct task dataset DeepPCB, consists six different types defects. Defect benefits both Among many group affordable, general-purpose, small-size very popular, referred as micro-PCBs. The those will facilitate process. called micro-PCB , includes 12 small Through comparative experiments, our system demonstrates its advantage tasks.