Motion Detection Index Testing
A set of applications built for testing the practical performance of an index built for storing motion detection data from surveillance video. The tests have only been built and tested on Mac OS X, but should work across platforms.
Programs
vid2diff.py
Take a video file as input, and create a set of grayscale difference frames with a given frame rate.
diffcompress.py
Create an index for a set of difference frames output from vid2diff.py. The index computes a set of histograms for the difference frames and compress the resulting set of histograms to the final index. Supports creating many variations of indices simultaneously (using different index parameters).
compareHistVideoTime.py
Compare the time spent answering a query on a video file, versus the time spent answering a query with an index built by diffcompress.py.
Data Sets
The folder datasets/ contains a number of files used for testing. There are three different scenarios encoded at three different resolutions each. These are useful for testing different possible use cases for surveillance video cameras.
Dependencies
Stuff required to run diffcompress.py and vid2diff.py
Python Packages
Use e.g. pip package manager to install these packages:
zlib
bz2
pylzma
liblzma
lz4
snappy
pywt
numpy
matplotlib
opencv
You must also install OpenCV. On OS X you can do so with Brew:
brew update
brew tap homebrew/science
brew install opencv
add to ~/.profile: export PYTHONPATH="/usr/local/lib/python2.7/site-packages/:$PYTHONPATH"
Command Line Packages
Requires following command line programs.
snappy
7z