Workpiece recognition based on improved Tiny-YOLOv3 algorithm

For the Tiny-YOLOv3 algorithm, the problem of high missed detection rate in real-time detection of workpiece recognition, Based on the improvement of Tiny-Yolov3, this paper realizes faster and more accurate recognition of small workpieces. The Tiny-Yolov3 feature extraction network adds three network modules, namely the SPP structure, the SE module and the Ghost module, and uses a convolutional layer to replace the pooling layer. Therefore, the average accuracy, accuracy and size of the network model of the improved network structure have been significantly improved. Experimental results show that the improved algorithm proposed in this paper can improve the efficiency of workpiece recognition and meet the requirements of real-time detection of embedded devices.