Contents

## **Preface XI**


**Space Representation of Motion Trajectories 101**  Shehzad Khalid

Preface

satisfy application specific requirements.

but still without reaching a final solution.

are introduced, analyzed and compared.

The book is organized into six chapters outlined as follows:

Surveillance systems have become an essential part in most establishments nowadays. There are many uses to these systems in national security, safety in public areas, flow control in crowded scenes, private safety and in providing special care for the aged and disabled. At the heart of any surveillance system are video cameras which have had their numbers significantly multiply over the last decade, thanks to advances in digital networks and automated video processing. This has resulted in an abundance of available surveillance video which made the process of monitoring them by human operators not only outdated but also practically infeasible. Several methods have been developed to automate the detection and reporting of scenes, events and subjects that

The purpose of this book is to collect recent advances in select areas of video surveillance. Research in that area usually combines results from machine learning, artificial intelligence, software engineering, stochastic modeling, signal processing in addition to pattern recognition and digital image/video processing. Solving problems related to video surveillance often requires the reconciliation between several contradicting objectives. This makes it a challenging task and also an open research area where novel solutions are continually presented to overcome earlier shortcomings

*Chapter 1* addresses the challenges that face surveillance applications due to the increased availability of visual data to be processed. As a case study, the problem of visual tracking is presented, its classical techniques are described and the difficulties that these techniques have to deal with due to increased visual data are illustrated. The emerging theory of compressive sensing is then introduced as a solution to these challenges, applying it to the successive stages of object tracking. Unlike the mathematical oriented approach used earlier, *Chapter 2* presents a software engineering approach to the problem of tracking. In particular, the technology of mobile agents is applied to the problem of tracking objects as they move between the fields of view of several cameras. The challenge here is to recover and maintain the identities of targets lost by the system. Neighborhood node determination techniques
