合成孔径声纳图像阴影增强方法研究

合成孔径声纳成像时，需要从不同位置采集目标的回波，从而造成目标的阴影被模糊。传统的声纳图像增强方法没有利用合成孔径声纳的成像机理，难以对阴影进行有效增强。从合成孔径声纳信号处理的角度，给出了固定焦距阴影增强（FFSE）算法的严格证明过程，导出了不同变换域下阴影增强的处理方式。对于多子阵合成孔径声纳，使用不均匀分离快速傅里叶变换（NSFFT）消除方位向的不均匀采样，然后在距离多普勒域进行阴影增强。分别对海上试验和湖上试验的数据进行了分析。处理结果表明，该方法能够显著增强合成孔径声纳图像的阴影，具有较强的实用价值，对于后续的图像分割和目标识别具有重要意义。

Research on Shadow Enhancement for Synthetic Aperture Sonar Images

Synthetic aperture sonar needs to collect the echoes from a target at different locations, which blurs the contour of shadow cast of targets. The conventional methods cannot obtain good effects because of lack in exploitation of synthetic aperture principle. A rigorous derivation of fixed focus shadow enhancement (FFSE) algorithm is presented, and time domain, frequency domain and range-Doppler domain are conducted. For synthetic aperture sonars, many receivers are used to increase efficiency. Non-uniform fast Fourier transform (NSFFT) is used to eliminate non-uniform sampling in azimuth direction, and then shadow enhancement can be applied in range-Doppler domain. The method is applied to lake-trial data set and sea-trial data set. Results show that the shadows of objects are enhanced evidently, which is great significance to image segmentation and target identification.