A Study On Infrared Dim Target Detection Method

Infrared (IR) imaging guidance system is one of the precision guided technologies developed by many counties of world, it also is a technique of high effectiveness and life cycle cost which enhances the missile might. The background of IR imaging guidance system is complex, such as sky, ground or sea. The task of the dissertation is to investigate the detect technique of infrared dim target embedded in complex background. The main contributions of the thesis are summarized as:Algorithm for edge detection and dim target detection method based on Wavelet- Hough Transform is proposed. The algorithm combines the Time-Frequency analysis with directionality of Hough Transform. The insensitivity of Wavelet Transform to all the other directions except the horizontal and vertical is avoided, and noise resistance is enhanced. Experiment for edge detection of three kinds of images and target detection in two kinds of backgrounds are done. Simulation experiment shows its efficiencies of noise resistance, directivity, and high performance in real time.General principles of one dimension and two dimension are presented in detail, respectively. To solve the difficulties in dim target detection, a new algorithm of automatic target detection based on wavelet lifting scheme and wavelet energy is proposed. Since this algorithm synthetically considers the inter-frame and single frame information, the detection result is more reliable and accurate.Dim target detection algorithm based on Empirical Mode Decomposition ( EMD ) is proposed. Bi-dimension EMD is introduced to decompose the original image. Then the residue is used to estimate the background, and remove the background, and then we get the target. Mathematical morphology filter is introduced to remove the noise. Theoretical analysis and experiment show the validity of our algorithm.