Improved Method of Garbage Classification Based on Deep Learning
Dong Wang, Zhongsheng Wang · 2021
Garbage classification is an important part of resource recycling and utilization, which can improve the utilization rate of resource recycling and reduce environmental pollution. Due to the wide variety of MSW and the lack of a unified standard for specific classification, most people will have difficulty in choosing MSW in practical operation. Traditional image classification methods have been difficult to meet the requirements. We can build an accurate classification model based on deep learning technology and use technical means to improve the living environment. In this paper, deep learning algorithm is used to identify and classify garbage, which can not only improve the efficiency of garbage classification, but also improve the accuracy of garbage classification. Mainly studies the garbage classification model based on CNN, methods of using the migration study selected suitable for network model of garbage classification in this paper, and then use the model integration methods such as increase the generalization ability of the model, finally using distillation technology will big model learned knowledge transferred to small in the model, the loss of lesser accuracy at the same time, improve efficiency and prediction accuracy of model prediction, To quickly and accurately identify the garbage type effect.