File size: 10,621 Bytes
d5f2c95 d9c7212 48116f7 17fe572 48116f7 17fe572 48116f7 d9c7212 d5f2c95 0a4c97a d5f2c95 0a4c97a d9c7212 d5f2c95 d9c7212 d5f2c95 0a4c97a d5f2c95 d9c7212 d5f2c95 d9c7212 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 |
---
license: mit
tags:
- remote-sensing
- computer-vision
- diffusion-models
- controlnet
- generative-model
- earth-observation
- open-vocabulary
- image-dataset
---
<p align="center">
<img src="assets/EarthSy.png" alt="Image" width="120">
</p>
<div align="center">
<h1 align="center"> EarthSynth: Generating Informative Earth Observation with Diffusion Models</h1>
<h4 align="center"><em>Jiancheng Pan*, Shiye Lei*, Yuqian Fu✉, Jiahao Li, Yanxing Liu</em></h4>
<h4 align="center"><em>Xiao He, Yuze Sun, Long Peng, Xiaomeng Huang✉ , Bo Zhao✉ </em></h4>
<p align="center">
<img src="assets/inst.png" alt="Image" width="400">
</p>
\* *Equal Contribution* Corresponding Author ✉
</div>
<p align="center">
<a href="https://arxiv.org/abs/2505.12108"><img src="https://img.shields.io/badge/Arxiv-2505.12108-b31b1b.svg?logo=arXiv"></a>
<!-- <a href="http://arxiv.org/abs/2408.09110"><img src="https://img.shields.io/badge/AAAI'25-Paper-blue"></a> -->
<a href="https://jianchengpan.space/EarthSynth-website/index.html"><img src="https://img.shields.io/badge/EarthSynth-Project_Page-<color>"></a>
<a href="https://huggingface.co/datasets/jaychempan/EarthSynth-180K"><img src="https://img.shields.io/badge/%F0%9F%A4%97%20Dataset-HuggingFace-yellow?style=flat&logo=hug"></a>
<a href="https://huggingface.co/jaychempan/EarthSynth"><img src="https://img.shields.io/badge/%F0%9F%A4%97%20Model-HuggingFace-yellow?style=flat&logo=hug"></a>
<a href="https://github.com/jaychempan/EarthSynth/blob/main/LICENSE"><img src="https://img.shields.io/badge/License-MIT-orange"></a>
</p>
<p align="center">
<a href="#news">News</a> |
<a href="#abstract">Abstract</a> |
<a href="#dataset">Dataset</a> |
<a href="#model">Model</a> |
<a href="#statement">Statement</a>
</p>
## Examples
A satellite image of road.
|  |  |  |  |  |
|---|---|---|---|---|
A satellite image of small vehicle.
|  |  |  |  |  |
|---|---|---|---|---|
A satellite image of tree. (Flood)
|  |  |  |  |  |
|---|---|---|---|---|
A satellite image of water.
|  |  |  |  |  |
|---|---|---|---|---|
A satellite image of baseball diamond, vehicle.
|  |  |  |  |  |
|---|---|---|---|---|
## TODO
- [ ] Release EarthSynth Training Code
- [x] Release EarthSynth Models to 🤗 HuggingFace
- [x] Release EarthSynth-180K Dataset to 🤗 HuggingFace
## News
- [2025/10/30] EarthSynth Models is uploaded to 🤗 [HuggingFace](https://huggingface.co/jaychempan/EarthSynth).
- [2025/8/7] EarthSynth-180K dataset is uploaded to 🤗 [HuggingFace](https://huggingface.co/datasets/jaychempan/EarthSynth-180K).
- [2025/5/20] Our paper of "EarthSynth: Generating Informative Earth Observation with Diffusion Models" is up on [arXiv](https://arxiv.org/abs/2505.12108).
## Abstract
Remote sensing image (RSI) interpretation typically faces challenges due to the scarcity of labeled data, which limits the performance of RSI interpretation tasks. To tackle this challenge, we propose **EarthSynth**, a diffusion-based generative foundation model that enables synthesizing multi-category, cross-satellite labeled Earth observation for downstream RSI interpretation tasks. To the best of our knowledge, EarthSynth is the first to explore multi-task generation for remote sensing, tackling the challenge of limited generalization in task-oriented synthesis for RSI interpretation. EarthSynth, trained on the EarthSynth-180K dataset, employs the Counterfactual Composition training strategy with a three-dimensional batch-sample selection mechanism to improve training data diversity and enhance category control. Furthermore, a rule-based method of R-Filter is proposed to filter more informative synthetic data for downstream tasks. We evaluate our EarthSynth on scene classification, object detection, and semantic segmentation in open-world scenarios. There are significant improvements in open-vocabulary understanding tasks, offering a practical solution for advancing RSI interpretation.
<p align="center">
<img src="assets/EarthSynth-FM.png" alt="Image" width="500">
</p>
## Dataset
EarthSynth-180K is derived from OEM, LoveDA, DeepGlobe, SAMRS, and LAE-1M datasets. It is further enhanced with mask and text prompt conditions, making it suitable for training foundation diffusion-based generative model. The EarthSynth-180K dataset is constructed using the Random Cropping and Category Augmentation strategies.
<p align="center">
<img src="assets/EarthSynth-180K-Map.png" alt="Image" width="400">
</p>
<p align="center">
<img src="assets/EarthSynth-180K.png" alt="Image" width="400">
</p>
### Data Preparation
We use category augmentation on each image to help the model better understand each category and allow more control over specific categories when generating images. This also helps improve the combination of samples in the batch-based CF-Comp strategy. If you want to train a remote sensing foundation generative model of your own, this step is not necessary. Here is the use of the category-augmentation method.
- Merge the split zip files and extract them
```
cat train.zip_part_* > train.zip
unzip train.zip
```
- Store the dataset in the following directory structure: `./data/EarthSynth-180K`
```
.(./data/EarthSynth-180K)
└── train
├── images
└── masks
```
- Run the category augmentation script:
```
python category-augmentation.py
```
After running, the directory will look like this:
```
..(./data/EarthSynth-180K)
└── train
├── category_images # Augmented single-category images
├── category_masks # Augmented single-category masks
├── images
├── masks
└── train.jsonl # JSONL file for training
```
## Model
### Environment Setup
The experimental environment is based on [`diffusers==0.30.3`](https://huggingface.co/docs/diffusers/v0.30.3/en/installation), and the installation environment references mmdetection's installation guide. You can refer to my environment `requirements.txt` if you encounter problems.
```
conda create -n earthsy python=3.8 -y
conda activate earthsy
pip install -r requirements.txt
git clone https://github.com/jaychempan/EarthSynth.git
cd diffusers
pip install -e ".[torch]"
```
### EarthSynth with CF-Comp
EarthSynth is trained with CF-Comp training strategy on real and unrealistic logical mixed data distribution, learns remote sensing pixel-level properties in multiple dimensions, and builds a unified process for conditional diffusion training and synthesis.
<p align="center">
<img src="assets/EarthSynth-Framwork.png" alt="Image" width="700">
</p>
### Train EarthSynth
This project is based on diffusers' ControlNet base structure, and the community is open for easy use and promotion. By modifying the config file of `train.sh` of the catalog `./diffusers/train/`.
```
cd diffusers/
bash train/train.sh
```
### Inference
Example inference using 🤗 HuggingFace pipeline:
```python
from diffusers import StableDiffusionControlNetPipeline, ControlNetModel
import torch
from PIL import Image
img = Image.open("./demo/control/mask.png")
controlnet = ControlNetModel.from_pretrained("jaychempan/EarthSynth")
pipe = StableDiffusionControlNetPipeline.from_pretrained("stable-diffusion-v1-5/stable-diffusion-v1-5", controlnet=controlnet)
pipe = pipe.to("cuda:0")
# generate image
generator = torch.manual_seed(10345340)
image = pipe(
"A satellite image of a storage tank",
generator=generator,
image=img,
).images[0]
image.save("generated_storage_tank.png")
```
Or you can infer locally:
```
python test.py --base_model path/to/stable-diffusion/ --controlnet_path path/to/earthsynth [--control_image_dir] [--output_dir] [--output_dir] [--category_txt_path] [--num_images]
e.g.
python test.py --base_model ./weights/stable-diffusion-v1-5/ --controlnet_path ./weights/EarthSynth/controlnet --num_images 5 --control_image_dir ./demo/control/ --category_txt_path ./demo/class.txt --output_dir ./outputs
```
### Training Data Generation
<p align="center">
<img src="assets/Vis.png" alt="Image" width="300">
</p>
### Acknowledgement
This project references and uses the following open-source models and datasets.
#### Related Open Source Models
- [Diffusers](https://github.com/huggingface/diffusers)
- [ControlNet](https://github.com/lllyasviel/ControlNet)
- [MM-Grounding-DINO](https://github.com/open-mmlab/mmdetection/blob/main/configs/mm_grounding_dino/README.md)
- [CLIP](https://github.com/openai/CLIP)
- [GSNet](https://github.com/yecy749/GSNet)
#### Related Open Source Datasets
- [OpenEarthMap](https://open-earth-map.org/overview_oem.html)
- [LoveDA](https://github.com/Junjue-Wang/LoveDA?tab=readme-ov-file)
- [DeepGlobe](http://deepglobe.org/)
- [SAMRS](https://github.com/ViTAE-Transformer/SAMRS)
- [LAE-1M](https://github.com/jaychempan/LAE-DINO)
### Citation
If you are interested in the following work or want to use our dataset, please cite the following paper.
```
@misc{pan2025earthsynthgeneratinginformativeearth,
title={EarthSynth: Generating Informative Earth Observation with Diffusion Models},
author={Jiancheng Pan and Shiye Lei and Yuqian Fu and Jiahao Li and Yanxing Liu and Yuze Sun and Xiao He and Long Peng and Xiaomeng Huang and Bo Zhao},
year={2025},
eprint={2505.12108},
archivePrefix={arXiv},
primaryClass={cs.CV},
url={https://arxiv.org/abs/2505.12108},
}
```
|