This code is based on the mmsegmentation framework. Please refer to the official document : https://mmsegmentation.readthedocs.io/zh-cn/latest/ to complete the configuration of the basic environment.
1.The data set is prepared according to the official document. If the DeepGlobe and CHN6 data sets are used, the configuration process of the data set can be omitted. This project has been implemented, and only the data set path used by the model needs to be modified according to the document.
2.The configuration file of this model is located in the mm Segmentation-main \ configs \ efemod folder, and the required configuration file is selected. According to the official document, the training can be started.