An Efficient Defect Detection System for Printed Circuit Boards with Edge-Cloud Fusion Computing

Many intelligent methods have been proposed and applied in the field of autonomous manufacturing inspection. These advanced algorithms with high requirements on computing power and network may lead to time delay, high cost and energy consumption in practical applications with massive data to be processed. We carry out an efficient defect detection system in an end-edge-cloud architecture with the concept of edge computing to process the big data quickly and effectively. A branchy deep learning model with early exit capability of inference is proposed to detect the category and location of the defect in printed circuit boards. We offload part of the computing tasks to the edge nodes by segmenting and deploying the DL model. Therefore, our system has high detection efficiency and makes real-time defect detection possible.