Back to datasets
Dataset assetOpen Source CommunityTerrain AnalysisFarmland Classification
Fine-Grained Farmland Dataset (FGFD)
We have developed an innovative dataset that covers farmland of various types across China, taking terrain variation into account.
Source
github
Created
Nov 27, 2024
Updated
Dec 9, 2024
Signals
709 views
Availability
Linked source ready
Overview
Dataset description and usage context
DBBANet
🌍 Fine‑Grained Farmland Dataset (FGFD)
- Description: This dataset comprises diverse farmlands across different terrain types in China.
- Annotations: Labels distinguish farmland without crops (red) from farmland with crops (green); in practice the focus is on separating farmland from non‑farmland.
- Download link: Baidu Cloud Download
- Extraction code:
FGFD
- Extraction code:
🏋️‍♀️ Training Methods
- Baseline Methods: A suite of comparison methods is provided for establishing baselines.
| Method | Description |
|---|---|
| UNet | UNet with ResNet‑50 encoder for segmentation. |
| DeeplabV3+ | DeepLabV3+ with ResNet‑50 encoder for segmentation. |
| PSPNet | Pyramid Scene Parsing Network for semantic segmentation. |
| HRNet | High‑Resolution Network for fine‑grained segmentation. |
| ABCNet | Attention Bilateral Context Network for efficient semantic segmentation. |
| CMTFNet | Fusion of CNN and multi‑scale Transformer for semantic segmentation. |
| MCCANet | Multi‑scale Channel‑Cross Attention Network with boundary supervision for semantic segmentation. |
| CGNet | Context Guided Network for efficient segmentation. |
| DenseASPP | Densely connected Atrous Spatial Pyramid Pooling network. |
| ENet | Efficient Neural Network for real‑time semantic segmentation. |
| SegNet | Encoder‑decoder network for pixel‑level classification. |
| BuildFormer | Model dedicated to building segmentation tasks. |
| UANet | Uncertainty‑aware network using ResNet‑50 for segmentation. |
| DSNet | Local‑global dual‑stream network for segmentation. |
| UNetFormer | UNet‑based model combined with Transformer layers. |
| DBBANet | Dual‑branch Boundary‑Aware Network for segmentation. |
- Training Steps:
- Set hyper‑parameters.
- Run the following command:
python train.py --batchsize 32 --model_name DBBANet --gpu_id 0
🧪 Testing Method
- Testing Steps:
- Ensure the model is properly trained and paths are correct.
- Run the following command:
python test.py --model_name DBBANet --batchsize 32
Need downstream help?
Pair the dataset with AI analysis and content workflows.
Once the source passes your review, move straight into summarization, transformation, report drafting, or presentation generation with the JuheAI toolchain.