FAST-Dynamic-Vision

FAST-Dynamic-Vision is a system for detection and tracking dynamic objects with event and depth sensing.

0. Overview

FAST-Dynamic-Vision is a detection and trajectory estimation algorithm based on event and depth camera.

Related Paper: FAST-Dynamic-Vision: Detection and Tracking Dynamic Objects with Event and Depth Sensing, Botao He, Haojia Li, Siyuan Wu, Dong Wang, Zhiwei Zhang, Qianli Dong, Chao Xu, Fei Gao

Video Links: YouTube, Bilibili

1. File Structure

• event-detector: Key modules with event processing, depth estimation, motion compensation and object detection.
• bullet_traj: modules for trajectory estimation in 3D world frame.

2. Setup

Requirement: Ubuntu 18.04 with ros-desktop-full installation; Ceres Solver; OpenCV 3; opencv_contrib

NOTION: If you are using Ubuntu 20.04 and failed to build this project with some linking error, please edit CMakeLists.txt and specify OpenCV 4 .

Step 1: Installation

sudo apt install libeigen3-dev build-essential libopencv-contrib-dev

Step 2: Clone the thie repo

git clone https://github.com/ZJU-FAST-Lab/FAST-Dynamic-Vision.git --branch=main

Step 3: build this project

cd FAST-Dynamic-Vision
catkin_make

3. Run a simple demo

Please clone branch dataset to use our demo dataset. It was recorded via DVXplorer, which provides 640x480 resolution.

3.1 Event-based motion compensation and object detection

This demo shows the performance of object detection and tracking algorithms.

source devel/setup.bash
roslaunch detector detection.launch

3.2 Moving ball trajectory estimation under motion-cap system

This demo shows detect and estimate the 3D trajectory of a throwing ball utilizing event and depth camera. The ground truth of the ball trajectory is captured by Vicon Motion Capture system.

source devel/setup.bash
roslaunch bullet_traj_est demo_traj_est_rviz.launch

4. Licence

The source code is released under GPLv3 license.

5. Maintaince

For any technical issues, please contact Haojia Li([email protected]), Botao He([email protected]), Siyuan Wu ([email protected]), or Fei GAO ([email protected]).

For commercial inquiries, please contact Fei GAO ([email protected]).