IMAGE PROCESSING, TRACKING AND EVALUATION TECHNIQUES FOR THERMAL IMAGING SYSTEMS

Abstract

In the past 40 years, thermal imaging or infrared imaging have advanced as essential technologies for armed forces. Since World War II, thermal imaging technology has been crucial for military operations. They function in the electromagnetic spectrum's infrared (IR) region. The essence of modern combat, especially homeland security, has been completely altered by IR technology. It allows the operator to see and acquire targets in total darkness and inclement weather. In comparison to other categories of night vision devices based on image intensification, thermal imaging systems have a number of benefits. The first difference between the two is that it may be used in full darkness, whereas image intensifiers need some kind of ambient light, even if it's only starlight. Systems that use thermal imaging may identify things at far greater distances than those that use image intensification. Unlike Radar, they are passive sensors and cannot be detected by enemy. Moreover, they have the ability to view even in mist, mild fog and rain. In view of the advantages mentioned above, thermal imaging systems are extensively used by defence forces for every platform, from airborne to small weapon systems. The development of thermal imaging systems based on high performance advanced infrared detectors requires an effective IR image processing module for high quality image. Integrated processing modules such as automatic target detection, tracking and track optimization for real time IR imaging is required for effective engagement of target from armoured platforms. Moreover, IR image processing should also address blooming issues faced during operation of thermal imagers on-board high power laser weapons for counter-drone systems. Worldwide, electro-optical industries are producing thousands of thermal imaging systems. Evaluation of a thermal imaging system is necessary before handing it over to a user. Evaluation of thermal imaging system performance is a tedious and complex task, if done through conventional evaluation technique. So, there is a need to devise new objective evaluation technique which is fast and accurate. Thus, the development of new and next generation thermal imaging system calls for the incorporation of an upgraded and efficient image processing algorithm, automatic target detection, tracking and objective evaluation technique for quick design validation. The goal of this research work is to thoroughly examine the methods for infrared object detection and tracking, image enhancement, image anti-blooming and evaluating the performance of thermal imaging system objectively. Following are four main objectives of the thesis:- a) Infrared image enhancement for high quality image b) Infrared object detection and tracking of target c) Anti-blooming technique in counter-drone system d) Performance evaluation of thermal imaging system 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. This thesis begins with an introduction to infrared imaging technology, design challenges, and the motivation of taking up the specific problem in Chapter 1. Chapter 2 presents an extensive literature survey on the latest developments in IR image enhancement techniques, IR object detection, tracking, anti-blooming methods available and performance evaluation techniques for Thermal imaging systems. Chapter 3 presents an edge and detail enhancement (EDE) algorithm which is an IR image processing technique to address the issue of image degradation due to varying atmospheric conditions. The work is novel in terms of proposed EDE algorithm. Moreover, elaborate discussion on problems encountered by thermal imager in Indian tropical climatic conditions ranging from mountains to marine area is included in the chapter. A rich and diverse IR image database was recorded and used for validation of EDE algorithm. This is again an important contribution in this field as there are no literature available regarding thermal imager’s operation in Indian tropical conditions. Qualitative and quantitative evaluation alongwith comparison with latest IR enhancement algorithms have been reported. Also, further addition is in terms of suggested hardware implementation of EDE algorithm on FPGA based video processing module. Chapter 4 presents an effective IR object detection algorithm using modified-Local Binary Pattern (M-LBP) with Particle Swarm Optimization (PSO) integrated kalman filter for tracking. This combination of modified-LBP for detection, kalman filter for tracking and track optimization using PSO in infrared domain is novel. Although a number of references exist where kalman and PSO have been used together yet, probably, this is the first time where both of these are integrated with modified LBP to solve a challenging task of tracking in infrared video. All of these three algorithms have their distinct characteristics and keeping them on a common platform makes it a powerful tracker for infrared imagery. Chapter 5 presents background for counter-drone systems, problem of blooming faced in field operation of high power laser based counter-drone system and image processing solution for anti blooming technique. This work is novel as the proposed anti-blooming technique applied over IR videos of actual anti drone operation is not available in literature. Chapter 6 presents an objective evaluation technique of thermal imaging systems based on Auto-MRTD. The Minimum Resolvable Temperature Difference (MRTD) is the primary key parameter of an IR imaging system, because it describes the overall system performance including that of a human operator or observer. MRTD is defined as the minimum temperature difference above 300 K required by an observer viewing through the device to resolve a vertical four bar pattern of 7:1 aspect ratio. An auto-MRTD method for performance assessment of thermal imaging system is suggested for production facilities. The laboratory method of MRTD is time-consuming and a costly process. To overcome these limitations, an auto-MRTD method is proposed. It will be helpful for industries producing thermal imagers on mass scale. Chapter 7 concludes about the research provided in this thesis. It highlights the key findings and goes on to examine the potential future research avenues based on the thesis findings.