/  20
 
Pedestrian TrackingIndependent Study Spring 2009
Isaac CaseMay 13, 2009
Abstract
One ability of the human visual system is the ability to detect and follow motion.This is a critical skill in understanding the world around us as it helps us identifyobjects of importance and helps us identify change in our environment. This ability ishighly desired in computer vision systems. Tracking motion can be used to monitoractivity in an area. For this work, I have applied this idea of object tracking in anattempt to track pedestrian motion in video.
1 Introduction
The goal of this work was to track people as they walk through some frames of video. Itwas intended to only track the motion of pedestrians. All other motion should be ignored.Partial and full occlusion should be compensated for. It should not matter how many peopleare in the video, or how large a person is, in comparison to the size of the video frame. Someof these goals were met, however, others were not able to be implemented due to the timeconstraint of this research, i.e. only ten weeks.1
 
2 Approach
2.1 Overview
In order to accomplish this goal, it was necessary to break this down into smaller pieces. Thecurrent implementation is based on the idea of background subtraction. Given a background,we are able to determine the difference from the current frame and the background to find theforeground. The foreground should be all of the objects that are worth investigation. Witha given foreground, the next step would be to segment it into regions of interest referred toas blobs. Then, with a set of blobs, track the motion from frame to frame, matching theseblobs across a series of frames.
2.2 Implementation
The implementation of this system followed the same process as described in the overview.Each part is a separate algorithm, however all work together to produce the results given.It may be possible to replace a subsection without any effect on any of the other subsectionsand in that way are somewhat independent and could be researched independently.
2.2.1 Background Detection
The first operation was to determine the background for every frame. A very simple approachwas tested at first. That implementation was as simple as just taking the average value of aseries of frames as the background as in:
B
=
k
n
=1
n
k
where
B
is the found background,
k
is the number of frames to be analyzed and
n
isthe
n
th
frame. This assumes that the background fills the frame a majority of the time,2
 
the background is fairly static, and the value of 
k
must be large enough force any non-background object to become insignificant. Given those conditions, the results can be fairlygood, however, given any deviation from those conditions and the results worsen quickly ascan be seen in figure 1(a) (b)Figure 1: A Comparison of Averaged backgrounds where (a) meets the criteria for averagingand (b) does not. (b) is averaged over too few framesOne major issue with this system of background detections is that it does not allowthe background to change over time. If an object was moving, but then stops moving, itshould become part of the background instead of forever staying as a foreground object. Oneexample of this is a car that enters a scene and then parks. The car when moving may be of interest and should be considered part of the foreground, but when the car parks and stops, iteventually should be ignored again as it is no longer a moving object. An improvement overthe previous method was also implemented for this research. It is based on the same ideaof averaging multiple frames of video, but instead of averaging every pixel for every frame,it only includes those pixels for which there is little motion over some range of frames. Itfollows a similar function, however has some important differences.
B
i,j
=
i,j
|
|
where
i,j
is the set of values for the pixel at position
i,j
in theseries of frames evaluated for which there is little change compared to the frame3

Share & Embed

More from this user

Add a Comment

Characters: ...