A waste image classification using convolutional neural networks and ensemble learning

Garbage classification is of great significance to environmental protection and resource recycling. Now many countries have passed laws related to garbage classification, defining different types of garbage. However, in the process of implementing these laws, it is found that correctly distinguishing different types of garbage is still a difficult task. In this paper, we will use a deep learning model to complete the task of garbage classification. Specifically, based on a publicly available image data set, a single convolutional neural network and the ensemble model based on these convolutional neural networks are compared for the classification performance. We found that the prediction results of the overall method are more accurate than a single neural network model, and among different ensemble approaches, random forest achieves the highest accuracy.