Permalink
Find file
Fetching contributors…
Cannot retrieve contributors at this time
19 lines (16 sloc) 1.07 KB
Copyright 2011-2012 Brandon L. Reiss
This software package contains implementations of two vehicle density
estimators designed for daytime and nighttime traffic density estimation. The
daytime estimator uses a Gaussian Mixture Model background subtraction
technique in order to label candidate pixels that correspond the vehicular
traffic. The nighttime technique uses a simple median filtering, averaging, and
morpholocial operations to perform bright object detection in a nighttime
sequence of traffic images. Both methods demonstrate potential to characterize
traffic density by the correlation of their output signals to the obeserved
traffic density in some image sequences. The nighttime algorithm is more robust
mostly due to the reduced complexity of the problem of finding bright
headlights in a mostly dark image.
The author of the software is Brandon L. Reiss, and he can be reached at
blr246@nyu.edu. All uses of the software are subject to the license agreement
prepended to each source file. A copy of the license is also included inside
the file called LICENSE.