maxibooksiyi/kcf_tracker_ros

Using kcf algorithm to track target under ROS environment.

About The Project

This project is based on a small sample of [DJI Onboard SDK](dji-sdk/Onboard-SDK: DJI Onboard SDK Official Repository (github.com)), the original version of this sample is implemented under linux environment rather than ros. This project uses KCF algorithms to track image target with bounding boxes.

Getting Started

Prerequisites

• Ros

You need to install ros first. Install instruction can be found at: http://wiki.ros.org/ROS/Installation. I just tested ROS kinetic version.

Installation

1. create your workspace

   mkdir -p you_ws/src
   cd you_ws/src

2. Clone the repo and complie

   git clone https://github.com/potBagMeat/kcf_tracker_ros.git
   cd ..
   catkin_make

3. download camera package of ros.

   sudo apt-get install ros-kinetic-usb-cam

Usage

Change camera

If you want to try other camera rather than usb camera. You should change ros topic name in target_tracking.cpp.

Firstly,

rostopic list

Find your topic and copy your topic name.

In target_tracking.cpp

change

image_transport::Subscriber imageSubscriber = it.subscribe("/usb_cam/image_raw",10,imageCallBack);

to

image_transport::Subscriber imageSubscriber = it.subscribe("you_topic_name",10,imageCallBack);

and don't forget to catkin_make.

Basic use

In order to run this project.

1. Run camera node.

roslaunch usb_cam usb_cam-test.launch

2. Run KCF tracker.

rosrun kcf_target_tracker kcf_target_tracker

and you will see

License

Distributed under the MIT License. See LICENSE for more information.

Contact

Xiaofeng Lin - [email protected]