Classification of features in open-pit mining areas based on deep learning and high-resolution remote sensing images

With the development of deep learning technology, the classification of high-resolution images has become a hot topic of current research, and the classification of features in mining areas is an important issue in the study of ecological development in mining areas.As deep learning can automatically identify and classify images by extracting a large number of historical image data patterns and features, this paper uses the U-Net model to carry out a study on the classification of features types in high-resolution opencast mining areas.Using the remote sensing image data of GF-2, the sample data set is outlined to extract the sample data features, and the classification model is trained to test the mining test set to explore the automatic recognition capability of deep learning on the high-resolution remote sensing images.The results show that the Precision, Recall, and F₁-score values of U-Net model for feature recognition in open-pit mine areas reach 0.86, 0.82, and 0.84, respectively, all of which are higher than those of the maximum likelihood method, random forest algorithm, and SVM method.U-Net network model based on deep learning can carry out effective automatic identification of features types in open-pit mining areas, which can provide technical support for features classification of high-resolution remote sensing image in open-pit mines, and effectively realize the ability of automatic identification and classification of properties in open-pit mining areas.The results of this paper can be used for the application of AI in the remote sensing classification of open-pit mines and the monitoring and restoration of the ecological environment of mining areas.