Repository files navigation

This is the demo part of team 14 presentation, in the course Video Communications (National Taiwan University, 2018 Spring).

slides

1. Download FFmpeg (version n4.0) inlib/ and apply patch

   $ git clone -b n4.0 https://github.com/FFmpeg/FFmpeg.git lib/FFmpeg-n4.0$ patch -p0< lib/ffmpeg-patch

2. Build FFmpeg with dependent packages

   $ sudo apt install yasm libx264-dev$cd lib/FFmpeg-n4.0$ ./configure --prefix=`pwd` --enable-gpl --enable-libx264$ make$cd -

3. Run the program, and open the output video

   $ sudo apt install python3-pip$ pip3 install numpy opencv-python opencv-contrib-python tqdm$ make start$ xdg-open output.mp4

The program will takedataset/animals_short.mp4 andoutput.mp4 as its input and output respectively.In order to change the input source video, you have to change theinput_video variable inmain.py.

             ╔═════ main.py ═════╗▶ frame ╾─┐  ║                   ║          ├──╫───╼ parser.py     ║▶  MVs  ╾─┘  ║         ┊         ║ (1-thread)  ║  homographier.py  ║             ║         ┊         ║             ║    detector.py ╾──╫──╼ output ▶             ║                   ║             ╚═══════════════════╝

• parser.py: Parse frame and motion vector informations fromFFmpeg. It's based onpython-mv with some alters for ffmpeg n4.0.
• homographier.py: Implement algorithms for estimating homography in each frame from UAV-video.
• detector.py: Use difference method to detect moving objects.

1. Referenced papers
   • A. Hafiane, K. Palaniappan and G. Seetharaman, "UAV-Video Registration Using Block-Based Features," IGARSS 2008 - 2008 IEEE International Geoscience and Remote Sensing Symposium, Boston, MA, 2008, pp. II-1104-II-1107.
   • Q. Wei, S. Lao and L. Bai, "Panorama Stitching, Moving Object Detection and Tracking in UAV Videos," 2017 International Conference on Vision, Image and Signal Processing (ICVISP), Osaka, 2017, pp. 46-50
2. Used libraries
   • FFmpeg, a collection of libraries and tools to process multimedia content such as audio, video, subtitles and related metadata.
   • Numpy, the fundamental package for scientific computing with Python.
   • OpenCV, the open source computer vision library.
   • radiant, an advanced image processing library, written primarily in Python.
   • tqdm: a fast, extensible progress bar for Python and CLI.
3. Test videos
   • Drone chasing animals | YouTube
   • VIVID Tracking Evaluation Web Site