Automation Lab, Sungkyunkwan University

GitHub Stats

Traffic Surveillance Dataset

1. Traffic Surveillance Data Generation capable of producing various environment record on road by using Carla.

2. Real-world traffic surveillance systems

I. Using CARLA

Building CARLA, Instruction, and samples

1. Building CARLA

Use git clone or download the project from CARLA Github.

Then follow the instruction at How to build on Linux or How to build on Windows.

The Linux build needs for an UE patch to solve some visualization issues regarding Vulkan. Those already working with a Linux build should insta

2. Instruction the patch and make the UE build again using the following commands.

# Download and install the UE patch  
cd ~/UnrealEngine_4.24
wget https://carla-releases.s3.eu-west-3.amazonaws.com/Linux/UE_Patch/430667-13636743-patch.txt ~/430667-13636743-patch.txt
patch --strip=4 < ~/430667-13636743-patch.txt

# Build UE
./Setup.sh && ./GenerateProjectFiles.sh && make

Please refer to INSTRUCTION.md for how to use.

3. Sample

Please go to this repository for Realistic-Traffic-Surveillance Generated Sample

II. Real-world System

We are conducting a research to develop a real-world traffic surveillance system.

Comprehensive, annotated dataset for object detection. This dataset consists of over 32 hours of real-world traffic surveillance data across 71 CCTV and an additional color cameras, spanning annual weather conditions (See Demo Videos). The UI for filtering scenes according to each attribute is provided in Releases: TSBOW-Filter-Scenes_v1.1 on Github repo.

Please go to TSBOW-dataset repository or TSBOW-website for more details.

III. Citation

If you find our work helpful for your research, please consider citing the following BibTeX entry.

1. Data Generation using Carla

@misc{AutoLab-Dataset-CARLA,
  author = {Automation Laboratory},
  license = {Apache-2.0},
  title = {Traffic Surveillance Dataset},
  howpublished = {https://github.com/SKKUAutoLab/ETSS-08-Data},
  year = {2025},
  note = {Data Generation using Carla}
}

2. Real-World System

@article{Huynh2026TSBOW, 
    title={TSBOW: Traffic Surveillance Benchmark for Occluded Vehicles Under Various Weather Conditions}, 
    volume={40}, 
    url={https://ojs.aaai.org/index.php/AAAI/article/view/37439}, 
    DOI={10.1609/aaai.v40i7.37439}, 
    number={7}, 
    journal={Proceedings of the AAAI Conference on Artificial Intelligence}, 
    author={Huynh, Ngoc Doan-Minh and Tran, Duong Nguyen-Ngoc and Pham, Long Hoang and Tran, Tai Huu-Phuong and Jeon, Hyung-Joon and Nguyen, Huy-Hung and Khac Vu, Duong and Jeon, Hyung-Min and Phan, Son Hong and Pham-Nam Ho, Quoc and Tran, Chi Dai and Khanh, Trinh Le Ba and Jeon, Jae Wook}, 
    year={2026}, 
    month={Mar.}, 
    pages={5239-5247} 
}

IV. License

Both the code and the weights pretrained on the COCO dataset are released under the Apache 2.0 license.