The Development The Integration Of Crowd's Movement Detection Into Surveillance Systems

Hidayah, Rahmalan and Nanna Suryana, Herman and Nor Azman, Abu and Aniza, Othman (2010) The Development The Integration Of Crowd's Movement Detection Into Surveillance Systems. Project Report. UTeM, Melaka, Malaysia. (Submitted)

PDF (Full Text) the_development_the_integration.pdf - Submitted Version Restricted to Registered users only Download (15MB)

Abstract

Being able to watch the CCTV system 24 hours for 7 days in order to prevent any crime or unwanted incident from occurring especially among the crowds, is a difficult and challenging job to do. This problem has led the integration of computer vision techniques into t he technology of CCTV system for a much better improvement in efficiency. Thus, the use of computer vision technique for surveillance application. especially in monitoring crowds has also increased.Initially, this study began with a review of research works related to subject areas that had great interest in crowd analysis. To identify significant parameters used in crowd analysis. a review was made on multidisciplinary aspects of crowd. including view from computer vision. In general. crowd analysis can be observed through three important tasks such as crowd density. crowd movement and crowd behaviour. According to Subbarao [58], human effectively perceive t he shape and mot ion of unfamiliar objects as they manage to observe t he changes of the objects· image. This shows t hat the changes of crowds can be measured using shape-based or motion-based techniques. However, Broggi et al. [11] stated that by applying the shape-based techniques. both moving and stationary pedestrians can be recognized. In addition. if the objects' shape presented in binary images. thus it can be described as the area of objects and then it can be used to describe the crowd density meanwhile. when there are changes between the shapes of objects. this can be used in measuring the movement of crowds. Thus. the motivation behind t his study is to investigate a shape-based technique as one of the computer vision techniques to be applied in crowd analysis. More specifically, for this study, the monitoring crowd task was focused for measuring the movement of crowds. Meanwhile. in terms of the usage of image data. sequence of images data were used for crowd movement experiment. As a result. there should also be an analysis of shape-based techniques applied for sequence of crowd images. before it has to be tested on real data of crowd. Since there were no standard of crowd images and there a re difficulties in capturing crowd images from public area. therefore the data used in this study are images of a crowd scene during a graduation day event in 2005 at University of Southampton, United Kingdom. simulating a surveillance application. T he following section describes the report outline according to chapter.

Actions (login required)

View Item