Real-time medicine box detection method based on YOLOv3 pruning network and embedded development board

The invention relates to a real-time medicine box detection method based on a YOLOv3 pruning network and an embedded development board. The method comprises the following steps: step 1, designing a YOLOv3 backbone network and a loss function; 2, collecting image data of the medicine boxes of all brands in a manual shooting mode; 3, making a medicine box data set and performing training; 4, performing model compression and acceleration calculation on the YOLOv3 through a pruning method based on a BN layer scaling factor gamma; 5, deploying the YOLOv3 compression model to a Nano embedded system,and carrying out model reasoning acceleration by using TensorRT; and step 6, carrying out real-time medicine box detection on the Nano by using a CSI camera. The invention is used for being deployedon an NVIDIA Jetson Nano embedded development board to carry out real-time medicine box detection, and the real-time performance of detection and the high efficiency of model operation are ensured while the detection precision