Infrared Signal Detection And Processing

In this thesis we will investigate IR signal detection and processing, where the study will be especially devoted to the detection of the moving Dim Point targets in IR image sequence algorithms and techniques, namely those used with modern staring infrared technology in performing the IRST function. To establish this goal we will study thoroughly every aspect related with the IR radiation, factors affecting the IR target detection. IR systems and the different techniques and algorithms used in the detection and processing.First we will study the electromagnetic spectrum and origin of the IR radiation, the effect of the atmosphere on the IR radiation propagation including the effect of every component of the atmosphere and the different codes that have been developed to represent this effect and propagation. Also we will study IR image and image sequence constitutes and its modeling and the IR scene radiation and its effect on the image contrast.Also we will study the IR detection system constitutes and categories starting from Electro-optical systems as general till we reach the IR detection systems and some of their military systems, also we will go through the probabilistic and statistical theories that are used in the case of random ergodic process, the theory of IR target detection and processing techniques and algorithms, the3D motion description and modeling and we will end with the study of different theories and algorithms in IR image sequence signal processing.In our proposals we will follow the concept of the track before detect (TBD) category, firstly because the detect before track (DBT) techniques and algorithms category were adequate for applications where the targets were bright compared with the background and it’s poor performance with dim targets in severe clutter the fact that the temporal behavior of the target and clutter was not used to its full extent, since spatial detection and thresholding were performed first. Secondly because the TBD techniques and algorithms category are a temporal based algorithm which uses multiple frames to incorporate temporal as well as spatial information.As a final goal of this thesis we will introduce an Adaptive Recursive Statistical Algorithm for IR Clutter Rejection built on the temporal behavior of IR clutter, noise and target based on a single-pixel basis. Where we will first create a3D cubes that represent the temporal behavior of each pixel, and then we will use this3D cubes to statistically check the presence of clutters and perform and effective rejection to all static signal (clutter) and keeping non-static weak signals, which consequently increased the signal to clutter ratio (SCR) for the IR image sequence under processing. A test will be performed on the algorithm to show its efficiency in the IR-clutter rejection.Finally we will introduce a new approach in processing the Infrared image sequence that have moving Dim Point targets detection by transformation of the IR-image sequence into4- vectors for each frame in the sequence. Where these4-vectors represent that standard deviation of columns, vectors and two diagonals of each image frame pixels, thus any point on the IR-image can be presented by two different pairs of coordinate either axial or diagonal. As an application to this new approach we will develop a new recursive algorithm to measure the image background Baseline which help in assigning an adaptive frame by frame threshold. And also we will develop an efficient adaptive recursive statistical spatio-temporal algorithm for detecting and tracking the moving dim point targets by correlating the four vectors and assuring that the target can be sensed in at least three of the four vectors. Also a test will be performed on the proposed approach and algorithms to prove their detection performance and ability to adapt and response to different dim small targets motion patterns.