Event-Aided Time-to-Collision Estimation for Autonomous Driving (ECCV2024)

Welcome to the official repository of Event-Aided Time-to-Collision Estimation for Autonomous Driving, an open-source project presented at ECCV 2024. This repository contains the implementation and relevant materials for our paper on time-to-collision estimation using event-based cameras. The project page is available here.

Overview

This project focuses on leveraging neuromorphic event-based cameras for accurate Time-to-Collision (TTC) estimation, especially in high-speed dynamic scenarios. Event cameras offer several advantages over traditional frame-based cameras, such as low latency, high temporal resolution, and efficient handling of high-contrast environments, making them ideal for collision prediction tasks.

Installation

Please follow the steps below to set up the project:

Prerequisites

• Windows 10/11 + MATLAB R2022a

Setup

1. Clone the repository:

   git clone https://github.com/NAIL-HNU/event_aided_ttc.git

2. Download the dataset:

   • The dataset can be downloaded from this link.
   • Place the downloaded dataset FCWD in the Datasets directory as follows:

     strttc_matlab_ws
     ├── Datasets
     │   └── FCWD
     └── matlab_code
         ├── main
         │   ├── function
         │   │   └── calculateSTRTTC.m
         │   └── main_FCWD.m
         └── utils
             ├── Find_nearest_index.m
             ├── GetValidPointonNLTS.m
             ├── ......
             └── strttc_optimize.m
     

Usage

MATLAB Code Instructions

Before running the MATLAB code, please modify the following paths in main_FCWD.m:

• root_path = "C:\Users\lijin\Downloads\test\strttc_matlab_ws\";
• datasetRootDir = 'C:\Users\lijin\Downloads\test\strttc_matlab_ws\Datasets\FCWD\20240302';

Once the paths are set, you can run the code by executing main_FCWD.m.

Project Page

For more detailed information, examples, and visual results, please visit our project page.

Results

Our method achieves state-of-the-art performance in several TTC estimation benchmarks. Detailed quantitative and qualitative results can be found on the project page.

Video Demonstration

Check out our video demonstration on YouTube:

Citation

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

@misc{li2024eventaidedtimetocollisionestimationautonomous,
    title         = {Event-Aided Time-to-Collision Estimation for Autonomous Driving},
    author        = {Jinghang Li and Bangyan Liao and Xiuyuan LU and Peidong Liu and Shaojie Shen and Yi Zhou},
    year          = 2024,
    eprint        = {2407.07324},
    archiveprefix = {arXiv},
    primaryclass  = {cs.CV}
}

License

This project is licensed under the MIT License. See the LICENSE file for details.

Acknowledgments

We thank Javier Hidalgo-Carrió and Davide Scaramuzza for releasing the design of the Beamsplitter in Event-aided Direct Sparse Odometry (EDS), based on which we build our FCW system. We also thank Dr. Yi Yu for proof reading. This work was supported by the National Key Research and Development Project of China under Grant 2023YFB4706600.

Contact

For any questions or inquiries, please contact us at:

• JinghangLi: [email protected]

Or open an issue in this repository.