MCCG

This repository contains the dataset link and the code for our paper MCCG: A ConvNeXt-based Multiple-Classifier Method for Cross-view Geo-localization, IEEE Transactions on Circuits and Systems for Video Technology. Thank you for your kindly attention.

Requirement

1. Download the University-1652 dataset
2. Download the SUES-200 dataset
3. Configuring the environment
   • First you need to configure the torch and torchision from the pytorch website

   • pip install -r requirement.txt

About dataset

The organization of the dataset.

More detailed about Univetsity-1652 dataset structure:

├── University-1652/
│   ├── train/
│       ├── drone/                   /* drone-view training images 
│           ├── 0001
│           ├── 0002
│           ...
│       ├── street/                  /* street-view training images 
│       ├── satellite/               /* satellite-view training images       
│       ├── google/                  /* noisy street-view training images (collected from Google Image)
│   ├── test/
│       ├── query_drone/  
│       ├── gallery_drone/  
│       ├── query_street/  
│       ├── gallery_street/ 
│       ├── query_satellite/  
│       ├── gallery_satellite/ 
│       ├── 4K_drone/

More detailed about SUES-200 dataset structure:

├── SUES-200/
│   ├── train/
│       ├── 150/
│           ├── drone/                   /* drone-view training images 
│               ├── 0001
│               ├── 0002
│               ...
│           ├── satellite/               /* satellite-view training images       
│       ├── 200/                  
│       ├── 250/  
│       ├── 300/  
│   ├── test/
│       ├── 150/  
│           ├── query_drone/  
│           ├── gallery_drone/  
│           ├── query_satellite/  
│           ├── gallery_satellite/ 
│       ├── 200/  
│       ├── 250/  
│       ├── 300/  

Train and Test

We provide scripts to complete MCCG training and testing

• Change the data_dir and test_dir paths in run.sh and then run:

bash run.sh

Citation

@ARTICLE{Shen2024MCCG,
  author={Shen, Tianrui and Wei, Yingmei and Kang, Lai and Wan, Shanshan and Yang, Yee-Hong},
  journal={IEEE Transactions on Circuits and Systems for Video Technology}, 
  title={MCCG: A ConvNeXt-Based Multiple-Classifier Method for Cross-View Geo-Localization}, 
  year={2024},
  volume={34},
  number={3},
  pages={1456-1468},
  keywords={Feature extraction;Drones;Task analysis;Image segmentation;Semantics;Satellites;Data mining;Cross-view;ConvNeXt;image retrieval;multiple feature representation},
  doi={10.1109/TCSVT.2023.3296074}}