Video surveillance based analysis of vehicles and pedestrians in intelligent traffic systems under complex environment

Abstract

iii newlineABSTRACT newlineAs there is an increase in the number of vehicles, Intelligent newlineTransportation Systems (ITS) becomes crucial and necessary in today s newlineovercrowded world. The major problems in transportation such as flexibility and newlinesecurity of vehicles are considered more serious in metropolitans. ITS mainly newlinefocus to attain secure traffic conditions and to increase safety. ITS applications newlinecan be used in different fields such as the management of traffic and intelligent newlinevehicles. newlineIn this thesis, the first work focuses on estimating the threshold value newlinefor the background subtraction method. Second, we work on subtracting the newlineforeground objects from the background, using the running-gaussian mixture, newlineframe-differencing, and three-frame differencing using logical AND and logical newlineOR method. Third, we design a new dimensionality-reduction method for newlineconstructing the low-dimensional data from high-dimensional data. In the fourth newlinemodule, we use a new L-shape feature descriptor from a 3D bounding box to newlineclassify the vehicles and pedestrians in traffic video. In the last module, we track newlinethe moving objects by identifying the four features from the 3D bounding box and newline