Siclelobe Suppression Algorithm Of Synthetic Aperture Sonar Imaging

Synthetic Aperture Sonar (SAS) is a high resolution underwater imaging sonar, which is similar to Synthetic Aperture Radar referred as SAR from principle to application. Imaging sonar obtains high range resolution by using the relative motion between the sonar and target to form large aperture, and obtains high azimuth resolution by transmitting linear FM signals with large time and bandwidth, and adopting the pulse compression technology. So SAS achieves the realization of the 2D high resolution imaging.The first, this paper introduces the imaging theory of SAS, including Linear FM and the technology of impulse compression. The second, this paper presents RDA and CSA as SAS image with high resolution. RDA (Range-Doppler Algorithm) is the most intuitive, basic and classical SAS imaging algorithm. CSA (Chirp Scaling Algorithm) is a improved algorithm of RDA, it keeps the phase accuracy of SAS image, so it has well effect of imaging. Many SAS imaging algorithms such as RDA and CSA adopt impulse compression, it need 2-D FFT and make time-domain signal look like Sinc function in 2-D because frequency signal in 2-D has finite width, so the interference of sidelobe is serious bad, the characteristic of impulse compression is existing a main lobe and sidelobe extended along each direction, The influence of sidelobe can make the image target not clear, and lower the image quality, so this paper designs improved algorithms to suppress sidelobe of images produced by both RDA and CSA imaging algorithms, then improve the quality of imaging, this paper’s four works are showed below:1. Proposing four evaluation parameter of SAS image quality:(1)Range error(RE), (2)Azimuth error(AE), (3)Peak-Sidelobe Ratio(PSR), (4)Peak-Image Ratio(PIR). The first two are location parameter of target point, in this paper it can be used to evaluate the capability of SAS imaging algorithms for locating point targets, so it is an important aspect of the SAS imaging quality. The last two are parameters of sidelobe, it can evaluate the amount of sidelobe and suppression capability of the improved algorithm such as window, SVA, and their combination.2. Adopting a sidelobe suppression algorithm called windowed algorithm, it weighs the amplitude of original data in time domain, then analyzes and compares the effect of rectangular window, Hamming window, Hanning window, finally, we chose Hamming window because it has the best comprehensive performance, it suppressed the further sidelobe. Windowed algorithm has already achieved some effects on sidelobe suppression, it applies to RDA is called W-RDA, it applies to CSA is called W-CSA.3. Due to the limitation of the windowed algorithm which broaden the main lobe and lower the SAS image resolution, this paper introduces the spatially variant apodization, it deal the original plural image with different weighting functions, then selects minimum values as the results in different weighted image, SVA can make each point of the SAS image in the frequency domain be corresponding weighted in amplitude, which has better effect than the traditional window function in sidelobe suppression while maintains a good resolution of the main lobe, it suppress the sidelobe near the target. SVA applies to RDA is called SVA-RDA, it applies to CSA is called SVA-CSA.4. Proposing a combination of windowed algorithm and SVA innovatively through comparing their advantages and disadvantages, the combination removes both far and close sidelobe around the target, that’s to say, it overcomes the deficiency of the previous two and has a better result. The combination algorithm applies to RDA is called WS-RDA, it applies to CSA is called WS-CSA.