Research on the Recognition and Classification Method of Recyclable Garbage Based on Improved YOLOv7

According to the chaotic situation of urban waste management, recyclable waste can be identified and classified accurately. This paper proposed an improved YOLOv7 recyclable garbage recognition and classification algorithm. First, the Rtrash recyclable garbage recognition and classification data set was constructed, and then an improved YOLOv7 attention mechanism algorithm SIM-YOLOv7 was designed. Two SimAM parameterless attention modules were added to the YOLOv7 neck network. The attention mechanism assigns 3D attention weights to garbage feature graphs to enhance the feature extraction capability of the model. In addition, three kinds of YOLOV7 networks with improved attention mechanisms were designed for comparison experiments. The experimental results show that the MAP of the YOLOv7 model is slightly improved after the addition of the attention mechanism. The evaluation index value of SIM-YOLOv7 increased more than that of the YOLOv7 model with other attention mechanisms. The superiority of this method in the recognition and classification of recyclable garbage is further verified.