Change-Agent: Toward Interactive Comprehensive Remote Sensing Change Interpretation and Analysis

Chenyang Liu, Keyan Chen, Haotian Zhang, Zipeng Qi, Zhengxia Zou, and Zhenwei Shi*✉

Share us a ⭐ if you're interested in this repo

Official PyTorch implementation of the paper: "Change-Agent: Toward Interactive Comprehensive Remote Sensing Change Interpretation and Analysis" in [IEEE] (Accepted by IEEE TGRS 2024)

News

• 2024-06: The code is available.
• 2024-03: The paper is available.

Table of Contents

• LEVIR-MCI dataset
• Training of MCI model
• Construction of Change-Agent
• Citation

LEVIR-MCI dataset

• Download the LEVIR_MCI dataset: LEVIR-MCI (Available Now!).
• This dataset is an extension of our previously established LEVIR-CC dataset. It contains bi-temporal images as well as diverse change detection masks and descriptive sentences. It provides a crucial data foundation for exploring multi-task learning for change detection and change captioning.
  
  
  
  

Training of the multi-level change interpretation model

The overview of the MCI model:

Preparation

• Environment Installation:

  Step 1: Create a virtual environment named Multi_change_env and activate it.

  conda create -n Multi_change_env python=3.9
  conda activate Multi_change_env

  Step 2: Download or clone the repository.

  git clone https://github.com/Chen-Yang-Liu/Change-Agent.git
  cd ./Change-Agent/Multi_change

  Step 3: Install dependencies.

  pip install -r requirements.txt

• Download Dataset:

  Link: LEVIR-MCI. The data structure of LEVIR-MCI is organized as follows:

  ├─/DATA_PATH_ROOT/Levir-MCI-dataset/
          ├─LevirCCcaptions.json
          ├─images
               ├─train
               │  ├─A
               │  ├─B
               │  ├─label
               ├─val
               │  ├─A
               │  ├─B
               │  ├─label
               ├─test
               │  ├─A
               │  ├─B
               │  ├─label
  

  where folder A contains pre-phase images, folder B contains post-phase images, and folder label contains the change detection masks.

• Extract text files for the descriptions of each image pair in LEVIR-MCI:

  python preprocess_data.py
  

  After that, you can find some generated files in ./data/LEVIR_MCI/.

Train

Make sure you performed the data preparation above. Then, start training as follows:

python train.py --train_goal 2 --data_folder /DATA_PATH_ROOT/Levir-MCI-dataset/images --savepath ./models_ckpt/

Evaluate

python test.py --data_folder /DATA_PATH_ROOT/Levir-MCI-dataset/images --checkpoint {checkpoint_PATH}

We recommend training the model 5 times to get an average score.

Inference

Run inference to get started as follows:

python predict.py --imgA_path {imgA_path} --imgB_path {imgA_path} --mask_save_path ./CDmask.png

You can modify --checkpoint of Change_Perception.define_args() in predict.py. Then you can use your own model, of course, you also can download our pretrained model MCI_model.pth here: [Hugging face]. After that, put it in ./models_ckpt/.

Construction of Change-Agent

• Agent Installation:

  cd ./Change-Agent/lagent-main
  pip install -e .[all]

• Run Agent:

  cd into the Multi_change folder:

  cd ./Change-Agent/Multi_change

  (1) Run Agent Cli Demo:

  # You need to install streamlit first
  # pip install streamlit
  python try_chat.py

  (2) Run Agent Web Demo:

  # You need to install streamlit first
  # pip install streamlit
  streamlit run react_web_demo.py

  
  
  

Citation

If you find this paper useful in your research, please consider citing:

@ARTICLE{Liu_Change_Agent,
  author={Liu, Chenyang and Chen, Keyan and Zhang, Haotian and Qi, Zipeng and Zou, Zhengxia and Shi, Zhenwei},
  journal={IEEE Transactions on Geoscience and Remote Sensing}, 
  title={Change-Agent: Toward Interactive Comprehensive Remote Sensing Change Interpretation and Analysis}, 
  year={2024},
  volume={},
  number={},
  pages={1-1},
  keywords={Remote sensing;Feature extraction;Semantics;Transformers;Roads;Earth;Task analysis;Interactive Change-Agent;change captioning;change detection;multi-task learning;large language model},
  doi={10.1109/TGRS.2024.3425815}}

Acknowledgement

Thanks to the following repository:

RSICCformer; Chg2Cap; lagent

License

This repo is distributed under MIT License. The code can be used for academic purposes only.

Contact Us

If you have any other questions❓, please contact us in time 👬