Detection Technology of Assisted Identification Laparoscopic Surgical Instruments Based on YOLO

The trained YOLOv5s and YOLOv5-Lite-s object detection models were deployed to the Raspberry PI 5 platform installed with Linux system respectively, and the deep learning environment was built on the platform. At the same time, OpenVINO tool was used to accelerate the performance optimization, and a laparoscopic surgical instrument detection system based on camera input was constructed. Through in-depth analysis and comparison of these two models, the experimental results show that compared with the original YOLOv5s model, the YOLOv5-Lite-s model, while maintaining close detection accuracy, significantly reduces the computation and memory consumption, and greatly improves the detection speed. The YOLa-based assisted identification and detection technology for laparoscopic surgical instruments proposed in this paper and the lightweight YOLOv5-Lite-s model for laparoscopic surgical instrument detection system have comprehensive advantages of high detection efficiency, low resource consumption, strong adaptability and low cost, and are very suitable for real-time control scenarios of embedded platforms.