Based On The Fractional Domain Time-frequency Characteristics Of Sar Image Features Target Detection Method Research

Synthetic aperture radar (SAR) is a coherent imaging radar which works in the microwave-band. It has many advantages, such as high resolution, all-time, all-weather, large area imaging, so it has been widely used in many aspects of national economic and military fields. However, with the development of SAR imaging technology in recent years, the SAR image acquisition technology has significantly been ahead of SAR image processing technology, therefore, it is particularly important for us to to make effective interpretation of SAR image. The main contents of this thesis are time-frequency features extraction in fractional domain of SAR image and ground target detection in SAR image based on these features.This thesis brings fractional short-time Fourier transform (FrSTFT) and fractional Wigner-Ville distribution (FrWD) in the field of SAR image target detection, and makes deep research in this field. The main research work and contributions this thesis has done are as follows:(1) The fractional domain time-frequency analysis is introduced into ground target detection on the two-dimensional SAR image signal processing from one-dimensional signal processing and this is innovative.(2) Fractional domain time-frequency analysis theory is researched in this thesis. Starting from the traditional time-frequency analysis methods, and based on the fractional Fourier transform, this thesis has made deep research in FrSTFT, FrWD, and the optimization of fractional domain time-frequency analysis, and also has made simulations and analysis on them.(3) A further study about the applications of fractional domain time-frequency analysis in SAR image analysis is done, including the display of fractional domain time-frequency data and the extraction of fractional domain time-frequency features of SAR image.(4) A ground target detection method of SAR images based on the fractional domain time-frequency features is proposed, which used fractional domain time-frequency features to detect targets. Then simulations have been done to validate the proposed method. And compared to the traditional methods, the proposed method has its advantages.