Research on Recyclable Garbage Classification Algorithm Based on Attention Mechanism

With the rapid development of global productivity levels, the problem of garbage disposal is getting more and more serious. Garbage classification is an important step to realize garbage reduction, harmlessness and resource utilization. With the increase in types and quantities of garbage, traditional garbage classification image algorithms can no longer meet the accuracy requirements of garbage identification. This paper proposes a ResNet18 convolutional neural network model based on the attention mechanism for the classification of recyclable garbage. The attention module is added after convolution, so that the model can pay more attention to the important information in the feature map. The model can automatically extract the characteristics of garbage for classification, including: glass, metal, plastic and paper. Experimental results show that the algorithm has an accuracy rate of 92 % in the classification of recyclable waste, which can effectively classify recyclable waste.