Gully-ERFNet: A novel lightweight deep learning model for extracting erosion gullies in the black soil region of Northeast China Original code

beYou need to prepare a folder to organize this database gully-erfnet ├─dataloader ├─dataset │ └─10m │   ├─biaoji │   ├─test │   ├─train │   ├─val │   └─yuantu ├─Loss fucation ├─networks ├─options ├─results ├─save_models_10m ├─test ├─train └─val Among them, dataloader is the function of loading data sets, dataset is the stored data sets, loss fucation is the loss function design, networks are the improved Gully-ERFNet network and erfent original network, options are the experimental setting parameters, results are the stored erosion gully extraction results, save_models is the model weight saving file, test, train, and eval are the training, testing, and verification functions of the model respectively.3.Other files include the result graphs of the two watersheds extracted by the model, the result graphs of the model comparisons, and the relevant environment configuration tables for model training.