多模式SAR成像及参数估计方法研究

发布时间：2018-06-04 22:08

本文选题：合成孔径雷达 + 多模式成像　；参考：《西安电子科技大学》2014年博士论文

【摘要】：合成孔径雷达(SAR)具有全天时、全天候、远作用距离、二维高分辨率及宽场景成像等特点,已在军事及民事领域得到广泛的应用。不同的应用,要求SAR图像具有不同的分辨率和场景覆盖范围,由此SAR发展出多种不同的工作模式,比如条带式、聚束式、滑动聚束式及TOPS SAR。同时,SAR灵活的波束指向又使其可以工作于正侧视、前斜视及前视等不同构型。由于不同模式不同构型的数据录取方式存在较大的差异,回波信号具有各自的特点,导致各种模式的信号处理方法不尽相同。本论文集中于多种SAR工作模式的信号处理,根据不同模式工作几何的特点,分析回波信号的特性,提出了合适的成像方法。具体内容如下:1.根据机载TOPS SAR工作模式波束扫描速度快、旋转角度大的特点,研究了一种新颖的全孔径成像方法。首先根据TWO-STEP思想进行方位预滤波,消除由天线方位旋转引起的多普勒频谱模糊问题,得到信号无模糊的二维频谱;其次结合非线性频调变标(NCS)算法完成距离单元徙动(RCM)校正及距离脉压;然后采用SPECAN的方法将数据聚焦在方位频域,避免聚焦在时域时出现的图像混叠问题;最后通过Chirp-Z变换,校正成像结果的几何形变。该算法不需要插值操作和坐标转换,因此运算量小、效率高。通过对仿真数据及某机载试验SAR系统录取的实测数据的处理,验证了方法的有效性。2.对广义极坐标格式算法(GPFA)进行了改进,使其可用于斜视TOPS SAR的数据处理。改进方法采用先线性距离走动校正后PFA插值的步骤。线性走动校正可以极大地降低距离向与方位向的耦合性,简化距离单元徙动(RCM)校正过程,并且有别于斜视聚束模式PFA插值常采用的坐标旋转思想,线性走动校正之后方位信号采样间隔依然是均匀的,这给方位向的插值带来了极大的便利。对于波束扫描及走动校正引起的多普勒调频率方位空变问题,采用方位非线性变标(NCS)的方法进行统一校正,大大提高了方位向聚焦深度,增加了可聚焦的场景范围。仿真和实测数据处理结果验证了所提方法的有效性。3.针对斜视多模式SAR的信号处理,提出了一种基于方位重采样的成像方法。斜视SAR回波信号由于存在线性走动,其距离向与方位向的耦合要显著得多。线性走动校正虽然可以极大地缓解这种耦合性,但同时也导致信号的方位空变性,使方位平移不变性不再成立。为了消除信号的方位空变性,我们提出了一种基于方位重采样的信号频谱处理方法。线性走动校正后,根据等效阵列的原理,在二维频域对信号频谱进行重采样,将斜视SAR信号的频谱等效为正侧视SAR。对于斜视波束指向SAR(BS-SAR),也即斜视聚束、斜视滑动聚束及斜视TOPS SAR,将方位重采样结合到信号频谱恢复的过程中,因此,由线性走动校正引入的方位空变性和由天线波束旋转引起的信号模糊都将被消除。经过线性走动校正及方位重采样后,斜视SAR数据就可以采用常规正侧视SAR成像算法完成聚焦。通过对仿真数据及实测数据的处理,验证了所提方法在多模式SAR成像中的有效性。4.将所提基于方位重采样的成像方法推广应用于斜视聚束式FMCW-SAR的数据处理。FMCW-SAR达到毫秒级的信号时宽导致常规算法推导中常用的“走-停-走”近似模型不再成立。针对此,并根据FMCW-SAR常采用Dechirp技术接收信号的特点,结合所提基于方位重采样的方法,我们提出了一种合适于斜视聚束式FMCW-SAR的全孔径成像方法。完成线性走动校正及方位卷积操作后,在距离多普勒域对雷达平台连续运动引起的多普勒频移进行补偿,避免在距离脉压结果中出现剩余的RCM。然后进行方位重采样,消除线性走动校正引起的方位空变性。最后执行RCM校正、二次距离脉压及方位匹配滤波,完成对数据的聚焦。此外,针对宽场景成像中RCM的距离空变问题,提出了一种基于子场景的校正方法。最后通过窄场景及宽场景的仿真实验对该方法进行了验证。5.提出一种采用周期性非均采样实现高分辨率宽测绘带的SAR成像体制。单通道星载SAR系统由于受最小天线面积的限制,高分辨率与宽测绘带构成一对矛盾。在所提的新体制中,采用发射正交波形及合适的系统PRF等方式,避免了距离模糊问题。利用非均匀的采样,并通过对采样序列的合理设计,使距离盲区均匀的分布在回波数据内,避开盲区的重叠现象;然后利用方位采样的周期性,构造出等效的多通道数据,并通过多通道解模糊的方法实现方位信号的频谱恢复。该体制可以实现星载单通道SAR系统高分辨宽测绘带成像,且不存在距离盲区问题。点目标的仿真实验验证了所提体制及方法的可行性及有效性。6.对多普勒调频率估计的平移相关(SAC)算法进行改进,使其适合用于聚束及滑动聚束SAR的全孔径成像处理。将距离Keystone变换引入到原始SAC算法中,消除了互相关峰值位置与目标距离间的耦合关系,从而去除了互相关函数幅度叠加时受到的聚焦深度的限制。对于因没有进行时域补零导致在调频率估计过程中出现的模糊问题,采用最小熵准则对模糊次数进行估计,并对调频率计算公式进行修正以得到正确的估计值。通过对仿真数据及某机载试验SAR系统录取的实测数据的处理,充分验证了改进SAC算法的优越性。
[Abstract]:Synthetic aperture radar (SAR) has the characteristics of all-weather, all-weather, far action distance, two-dimensional high resolution and wide scene imaging. It has been widely used in military and civil fields. Different applications require different resolution and scene coverage of SAR images, thus SAR has developed a variety of different working modes, such as strip type, Bunching, sliding cluster and TOPS SAR. simultaneously, the flexible beam direction of SAR enables it to work in different configurations such as Yu Zheng side view, front strabismus and forward view. Due to the large difference in the data admission mode of different modes and different configurations, the echo signal has its own characteristics, which leads to the different methods of signal processing in various modes. The thesis focuses on the signal processing of various SAR working modes. According to the characteristics of different modes of working geometry, the characteristics of the echo signal are analyzed and the appropriate imaging methods are proposed. The specific contents are as follows: 1. a novel full aperture imaging method is studied according to the characteristics of the fast beam scanning speed and the large angle of rotation in the working mode of the airborne TOPS SAR. Firstly, the azimuth pre filtering is carried out according to the TWO-STEP idea to eliminate the ambiguity of Doppler spectrum caused by the azimuth rotation of the antenna, get the two-dimensional spectrum without ambiguity, and then combine the nonlinear frequency modulation (NCS) algorithm to complete the distance unit migration (RCM) correction and the distance pulse pressure, and then use the SPECAN method to focus the data in the azimuth frequency. Domain, avoiding the problem of image aliasing when focusing on the time domain; finally, the geometric deformation of the imaging results is corrected by Chirp-Z transformation. The algorithm does not need interpolation operation and coordinate conversion, so the computation is small and efficient. The effectiveness of the method is verified by the processing of the simulated data and the measured data accepted by an airborne test SAR system,.2 The generalized polar coordinate format algorithm (GPFA) is improved so that it can be used for the data processing of the strabismus TOPS SAR. The improved method adopts the step of the PFA interpolation after the linear distance movement correction. The linear walk correction can greatly reduce the coupling between the distance to the azimuth, simplify the RCM correction process and be different from the deviation. The idea of coordinate rotation often used in the PFA interpolation of optical beam mode is used, and the sampling interval of azimuth signal is still uniform after linear motion correction, which brings great convenience to the azimuth interpolation. The azimuthal nonlinear scaling (NCS) method is used for the Doppler frequency azimuth change problem caused by beam scanning and moving correction. The unified correction greatly improves the orientation focus depth and increases the range of the focusable scene. The simulation and measured data processing results verify the effectiveness of the proposed method.3.. A method based on azimuth resampling is proposed for the signal processing of the strabismus multi mode SAR. The skew SAR echo signal has a linear motion and its distance due to the existence of a linear motion. The coupling of the off direction and the azimuth is much more significant. Although the linear motion correction can greatly alleviate this coupling, it also leads to the azimuth invariance of the signal, which makes the invariance of the azimuth shift no longer established. In order to eliminate the azimuth denaturation of the signal, we propose a method of signal spectrum processing based on azimuth resampling. After the correction, the signal spectrum is resampling in two-dimensional frequency domain according to the principle of the equivalent array, and the spectrum of the SAR signal is equivalent to the positive side view SAR. for the strabismus beam pointing to SAR (BS-SAR), that is, the strabismus bunching, the strabismus slipping bunching and the squint TOPS SAR, and the azimuth resampling in the process of the signal spectrum recovery. The azimuth denaturation introduced by linear motion correction and the signal blur caused by the antenna beam rotation are eliminated. After linear motion correction and azimuth resampling, the SAR data of the strabismus can be focused by the conventional positive side view SAR imaging algorithm. The effectiveness.4. in the SAR imaging makes the image method based on azimuth resampling applied to the data processing of the strabismus cluster FMCW-SAR data processing.FMCW-SAR to millisecond signal time, which leads to the usual "walk stop walk" approximation model, which is commonly used in the conventional algorithm deduction. With the feature of signal receiving, we propose a full aperture imaging method suitable for the strabismus bunching FMCW-SAR. After completing the linear walk correction and azimuth convolution operation, the Doppler shift caused by the continuous motion of the radar platform in the distance Doppler domain is compensated to avoid the distance pulse pressure junction. The remaining RCM. appears in the fruit and then carries out azimuth resampling to eliminate azimuth denaturation caused by linear motion correction. Finally, RCM correction, two distance pulse pressure and azimuth matching filtering are performed to focus on the data. In addition, a correction method based on the subscene based on the sub scene is proposed for the range space change problem of RCM in wide scene imaging. This method is verified by the simulation experiments of narrow scene and wide scene. A SAR imaging system with high resolution and wide swath with periodic non uniform sampling is proposed. The single channel spaceborne SAR system has a contradiction between the high resolution and the wide swath band because of the limitation of the minimum antenna area. In the proposed new system, the SAR system is used in the new system. The transmission of orthogonal waveforms and appropriate system PRF avoids the problem of distance blurring. Using non-uniform sampling and reasonable design of the sampling sequence, the distance blind area is evenly distributed in the echo data to avoid the overlap phenomenon of the blind area, and then the equivalent multi-channel data is constructed by using the periodicity of azimuth sampling, and it is passed. The spectrum recovery of azimuth signals is realized by the method of multichannel ambiguity resolution. The system can realize high resolution wide swath imaging of the satellite borne single channel SAR system with no distance blind area. The simulation experiment of point target verifies the feasibility and effectiveness of the proposed system and method and the translation correlation (SAC) algorithm for the estimation of Doppler frequency modulation rate (.6.). It is improved so that it is suitable for all aperture imaging processing for bunching and sliding cluster SAR. The distance Keystone transform is introduced into the original SAC algorithm to eliminate the coupling relationship between the position of the correlation peak value and the distance of the target, thus removing the limitation of the focus depth when the amplitude superposition of the cross correlation function is removed. For the reason that the time domain is not carried out in time domain. The fuzzy problem in the estimation of frequency modulation rate is caused by zero filling. The minimum entropy criterion is used to estimate the fuzzy number, and the formula of the frequency modulation is corrected to get the correct estimation value. The superiority of the improved SAC algorithm is fully verified by the processing of the simulated data and the measured data accepted by an airborne test SAR system.
【学位授予单位】：西安电子科技大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TN957.52

【相似文献】