PERFORMANCE EVALUATION OF VIDEO PROCESSING TECHNIQUES FOR ELECTRO-OPTICAL SURVEILLANCE SYSTEM

Abstract

An Electro-optical (EO) sensor operates by modifying the optical properties of a material by an electric field. EO sensors are totally passive, sensing energy emitted by or reflected off the target itself. Thus, it makes EO sensors ideal for surveillance application to acquire real-time video information of area of interest [1]. EO based surveillance system imply imaging sensors operating in the visible and infrared portions of the electromagnetic spectrum. EO sensors operating in visible band are used to exploit color data and high quality image, whereas, Infrared (IR) EO sensors benefit the surveillance task in night and under poor visibility condition [2]. Data collected through surveillance task is enormous. In order to reduce operator's load, it is important to perform majority of video processing operations through automated systems based on machine vision. The task of video processing is to improve pictorial information for human interpretation and to process the scene data automatically through machine vision modules [3]. To extract the useful information from the data being collected by EO sensors, the video processing techniques augment the need of pre-processing, information extraction and follow-up of extracted information. Thus, video processing for EO surveillance system has been a very active research topic in the last few years due to its growing importance in security, law enforcement, and military applications [4]. The aim of this research work is to explore the performance measures and evaluation techniques of video processing techniques for EO surveillance system. It encompasses in depth analysis of video processing techniques such as observation range enhancement, video tracking and comprehensive evaluation of EO system integrated with proposed video processing techniques. Based on the challenges presented in previous section for the implementation of video processing techniques and evaluation of EO system, there are four main objectives of the thesis. a) Analysis of video processing techniques for surveillance system b) Comparative evaluation of digital and hybrid techniques for observation range enhancement c) Efficient method for performance evaluation of EO tracking system d) Objective evaluation of EO imaging system This thesis primarily focuses towards the analysis, implementation and performance evaluation techniques for electro-optical surveillance systems. The thesis consists of seven chapters. Each chapter begins with a brief introduction to the concerned problem and its possible solution. The experimental results are summarized and concluded at the end of each chapter. A brief discussion about each chapter is presented below. This thesis begins with an introduction to electro-optical surveillance system, design challenges, and the motivation of taking up the specific problem in Chapter 1. Chapter 2 presents an extensive literature review on the latest development in analysis, real-time implementation and performance evaluation of applicable video processing techniques for EO based surveillance system. Chapter 3 presents performance analysis of electro-optical sensor, video processing for improved detection and tracking capability. Improvement in terms of object detection and robust tracking is demonstrated for maritime electro-optical surveillance system. EO surveillance system captures and extracts the information in real-time. Advancements in electro-optical sensor performance necessitate real-time processing of video processing techniques for effective surveillance. Chapter 4 presents indepth evaluation of real-time observation range enhancement techniques based on pixel manipulation, physical model and hybrid techniques employed for hassle free surveillance. The presented work includes qualitative and quantitative analysis of aforementioned techniques. Observation range enhancement of the order of 30% and image quality improvement in terms of universal image quality matrices is demonstrated. Chapter 5 presents an efficient method for evaluation and hardware-in-loop testing of electro-optical tracking system. The proposed method facilitates the efficient evaluation of a closed-loop video tracker by providing a virtually endless video sequence containing realistic visual attributes of potential objects. In chapter 6, quantitative evaluation of EO system performance consists of relevant performance evaluation indicators and an efficient evaluation mechanism is presented. The proposed work extensively addresses the performance evaluation of infrared sensor based imaging system. An objective minimum resolvable temperature difference (MRTD) measurement model has been proposed for the evaluation of a large number of imaging systems in the production line. The feasibility of proposed objective measurement has been demonstrated with accuracy greater than 95%. Chapter 7 concludes the research work, with conclusions based on the results obtained and major outcomes. Future scope of the work for further improvement is also deliberated in conclusion.