Automatic Classification of Remote Sensing Images of Landfill Sites Based on Deep Learning

China, as one of the world's largest producers of municipal solid waste, is faced with a huge amount of waste problem. Unregulated landfills not only lead to environmental hazards, but also may have landslide risks. Remote sensing technology has the characteristics of long-distance, non-contact and periodicity, which can make up for the shortage of crrent manual management methods. Therefore, this study will use deep learning combined with remote sensing images to achieve automatic classification of landfill images. The study uses CB04 and sentinel2A satellite images to construct a landfill remote sensing image dataset, and divides the dataset into training set, validation set and test set according to the ratio of 6:2:2. In this study, two classical models, Vgg and ResNet, are used to implement image classification on landfill remote sensing image datasets. The results show that (1) in this study, a method for automatic classification of remote sensing images of landfills is proposed, with a model accuracy of 86.76%, and (2) deep learning can effectively implement automatic landfill classification, and both Resnet and Vgg have shown good classification results., and (3) Resnet and Vgg can better perform the task of automatic landfill identification after introducing the optimization strategy of parameter migration. As shown above, deep learning has strong potential in the field of landfill image classification.