Open-Pit Mining Area Extraction from High-Resolution Remote Sensing Images Based on EMANet and FC-CRF

Rapid and accurate identification of open-pit mining areas is essential for guiding production planning assessing environmental impact. Remote sensing technology provides an effective means mine boundary identification. In this study, method delineating area from remote images proposed, which based on the deep learning model Expectation-Maximizing Attention Network (EMANet) fully connected conditional random field (FC-CRF) algorithm. First, ResNet-34 was applied as backbone network to obtain preliminary features. Second, EMA mechanism used enhance important information details in image. Finally, a postprocessing program FC-CRF introduced optimize initial prediction results. Meanwhile, extraction effect MobileNetV3, U-Net, convolutional (FCN), our were compared same data set areas. The advantage verified by visual graph results, accuracy evaluation index confusion matrix calculation. pixel (PA), mean intersection over union (MIoU), kappa 98.09%, 89.48%, 88.48%, respectively. results show that effectively identifies It practical significance complete task accurately comprehensively, can be management protection mines.