Multi-label Classification of Remote Sensing Image Based on Deep Learning on Embedded Device

In recent years, the resolution of satellite remote sensing images has been continuously improved. People can obtain more useful data and information from remote sensing images. Remote sensing images are widely used in various fields such as resource surveys and weather forecasting. The classification of remote sensing images is a very critical link in remote sensing image processing. This article proposes a multi-label classification method for remote sensing images based on deep learning for embedded platforms. We use residual neural network as the basic model and use transfer learning to reduces the cost of model training, besides this, we use batch normalization, image augmentation, and self-defined loss evaluation index to improve the accuracy and generalization ability of the model. Finally, the model is deployed on the embedded platform NVIDIA Jetson TX2 through tensorflow lite, the model performs well on the embedded platform and has good real-time performance.