Research on Real Time Object Detection Optimization Algorithm for Drill Model Parts Based on Deep Learning

A deep learning based real-time object detection algorithm for drilling die components is proposed to address the real-time online recognition problem in the assembly process of advanced aviation equipment. On the basis of YOLOv5 object detection algorithm, this algorithm studies a new cross layer connection method of deep learning algorithm, introduces SE attention mechanism and CBAM attention mechanism, optimizes the network architecture parameters of YOLOv5 deep learning model, and improves the overall performance of the optimized algorithm. Through comparative experiments, it can be found that the real-time object detection optimization algorithm for drilling mold components based on deep learning improves the recognition accuracy of drilling mold components, and has a high running speed, which can meet the real-time online object detection requirements of augmented reality assembly process.