3DInvNet

Implementation codes and datasets for the paper "3DInvNet: A Deep Learning-Based 3D Ground-Penetrating Radar Data Inversion" at http://dx.doi.org/10.1109/TGRS.2023.3275306.

1. The simulated and real measured datasets can be found at https://drive.google.com/drive/folders/12mG7lA8g8KX55KPHmP9y3n4XxFZjLp4W?usp=sharing.

2. Commands for training Inverter:
   python workspace_Inverter/trainInvNet.py
   --model InvNetModel_ff
   --lossfc MAE_loss
   --batch_size 4
   --lr 0.001
   --lr_decay 0.98
   --max_epoch 100
   --id MAE
   --train_data_path dataset/train/mask1
   --train_mask_path dataset/train/mask2
   --test_data_path dataset/test/mask1
   --test_mask_path dataset/test/mask2
   --model_path workspace_Inverter/exp/model/
   --visualization_path workspace_Inverter/exp/visual/
   --save_model

3. Commands for training Denoiser:
   python workspace_Denoiser/trainDenoisingNet.py
   --model DenoisingNetModel
   --lossfc MSE_loss
   --batch_size 4
   --lr 0.001
   --lr_decay 0.98
   --max_epoch 100
   --id MSE
   --train_data_path dataset/train/data
   --train_mask_path dataset/train/mask1
   --test_data_path dataset/test/data
   --test_mask_path dataset/test/mask1
   --model_path workspace_Denoiser/exp/model/
   --visualization_path workspace_Denoiser/exp/visual/
   --save_model

4. If any issues pls contact [email protected].