Vision-based Real-Time Aerial Object Localization and Tracking for UAV Sensing System

This repository contains the code (implemented in C++ on Linux) and dataset for our IEEE-Access-2017-paper.

If you find our code or project useful in your research, please cite:

@ARTICLE{Wu_Vision2017, 
author={Y. Wu and Y. Sui and G. Wang}, 
journal={IEEE Access}, 
title={Vision-Based Real-Time Aerial Object Localization and Tracking for UAV Sensing System}, 
year={2017}, 
volume={5},
pages={23969-23978},
doi={10.1109/ACCESS.2017.2764419}
}

Please also consider citing:

@inproceedings{zhang2015minimum,
title={Minimum barrier salient object detection at 80 fps},
author={Zhang, Jianming and Sclaroff, Stan and Lin, Zhe and Shen, Xiaohui and 
Price, Brian and Mech, Radomir},
booktitle={Proceedings of the IEEE International Conference on Computer Vision},
pages={1404--1412},
year={2015}
}

Prerequisite

OpenCV 2.4+

Cmake 2.8

Usage

1. Download our dataset and untar to the correct path (../../../dataset/), or you can use your own dataset.
2. go to /path/to/code/MBSPlusKF/, mkdir build
3. cd build
4. cmake ..
5. make
6. ./MBSPlusKF ../../../dataset/

(7. If want to save results: ./MBSPlusKF ../../../dataset/ ../../../Tracking_results/airplane_016/)