No abstract

The issue of garbage classification has aroused great concern for residents recently. The traditional garbage classification method has a low classification accuracy rate, high cost, and large personnel demand. With the increasing amount of garbage, it is difficult for traditional methods to effectively separate waste. Therefore, consider applying deep learning to the garbage classification problem. This paper uses two types of Convolutional Neural Networks (CNNs)--Inception V3 and Inception V4 to train Huawei's public garbage data set (Garbage Date) and establish a garbage classification model. After experiments, the classification results are compared and the performance of the model is tested. In this paper, by observing the changes in the accuracy rate and cross-entropy loss function of the two models on the training and test sets in the experiment, it is found that both models can obtain higher accuracy of garbage classification The network model using Inception V4 is more stable and accurate than the network model using Inception V3. The experimental results also show that this method can improve the accuracy of garbage classification in daily life and the efficiency of recyclable garbage collection.