A Research On Fourier Analysis Of The Infrared Image Marine Target Detection

Marine target detection has important practical significance in safeguarding marine rights and interests and national defense construction. Infrared detector has good advantages on invisibility, all-weather, electromagnetic interference immunity and so on, and it has been widely applied to monitor the sea condition. Because marine infrared images targets possess low signal to noise ratio, less information on target shape and texture, and other complex background factors such the wave, it brings a great difficulty for valid marine target detection in infrared images. Therefore, this paper deeply analyzes the characteristics of infrared images and infrared video images, particularly the imaging characteristics in the Fourier domain, and then systematically studies the marine infrared image target detection methods. The main research work is as follows:1) A marine target detection algorithm based on the multi-band pass filtering and fuzzy decision in semi-Fourier domain(SFD_Mb FFD) is proposed for marine infrared images. In the proposed algorithm, the characteristics of a target's data in semi-Fourier domain are analyzed, and the corresponding images perform multi-band filtering in semi-Fourier domain. Then multiple background-suppressed images filtered carry on the fuzzy comprehensive evaluation for obtaining an optimal background-suppressed image. Finally, this infrared image obtained is utilized for target detection. The experimental results demonstrate that the SFD_Mb FFD approach holds better background suppression effect for different types of marine targets, particularly for non-uniformity images. In addition, it can suppress marine background, highlight targets, detect marine target completely.2) A marine infrared video target detection algorithm based on background statistical modeling in Fourier domain(FD_SGM) is presented. It is first analyzed that the shortages of single Gauss background modeling and mixture Gaussian background modeling for detecting marine infrared video targets. Because the marine background possesses the stable wave pattern in spatial domain, its corresponding amplitude spectrum also has satiability in Fourier domain. Hence, the FD_SGM algorithm is established. The experiment illustrate that the model of marine background can realize marine target detection in many types of backgrounds.3) A marine infrared video target detection algorithm based on background statistical modeling on significant features in semi-Fourier domain(SFD_SGM) is developed. Because the amplitude spectrum of marine background possesses satiability in semi-Fourier domain, a background statistical modeling technology in semi-Fourier domain is proposed through an extended research on background statistical modeling in Fourier domain. Then according to the significant features of marine target in semi-Fourier,it can make fast acquisition of target region approximately. Finally in this region, marine targets are detected through marine background modeling in semi-Fourier domain. The experiment results indicate that the SFD_SGM method doesn't only detect marine targets efficiently in different sea condition, and it can reduce calculating complexity and improve detection efficiency.