This is an open-source implementation of paper: Map2DFusion: Real-time Incremental UAV Image Mosaicing based on Monocular SLAM.

Website : http://zhaoyong.adv-ci.com/map2dfusion/

Video : https://www.youtube.com/watch?v=-kSTDvGZ-YQ

PDF : http://zhaoyong.adv-ci.com/Data/map2dfusion/map2dfusion.pdf

• Download the latest code with:

  • Git:

      git clone https://github.com/zdzhaoyong/Map2DFusion
    

• OpenCV : sudo apt-get install libopencv-dev
• Qt : sudo apt-get install build-essential g++ libqt4-core libqt4-dev libqt4-gui qt4-doc qt4-designer
• QGLViewer : sudo apt-get install libqglviewer-dev libqglviewer2
• Boost : sudo apt-get install libboost1.48-all-dev
• GLEW : sudo apt-get install libglew-dev libglew1.10
• CUDA (optional) : see https://developer.nvidia.com/cuda-downloads

Warnning: Compilation with CUDA can be enabled after CUDA_PATH defined.

If you are using linux systems, it can be compiled with one command (tested on ubuntu 14.04):

cd Map2DFusion;make

Obtain the sample sequence and launch:

git clone https://github.com/zdzhaoyong/phantom3-village-kfs
./Map2DFusion DataPath=phantom3-village-kfs

More sequences can be downloaded at the NPU DroneMap Dataset.

If you have any issue compiling/running Map2DFusion or you would like to know anything about the code, please contact the authors:

 Yong Zhao -> zd5945@126.com