【作者】 江志红；

【导师】 皇甫堪；

【作者基本信息】 国防科学技术大学 ， 信息与通信工程， 2008， 博士

【摘要】 调频连续波合成孔径雷达(Synthetic Aperture Radar,SAR)是一种体积小、重量轻、造价低的高分辨成像雷达,非常适合无人机等小型飞行平台的遥感和侦察,具有广泛的应用前景。与常规的脉冲SAR相比,小型飞行平台应用背景下的调频连续波SAR实时成像主要面临四个问题:一是天线的连续运动不能忽略;二是系统采用去斜率的混频方式,采样频率很低;三是发射信号扫频非线性的影响必须考虑;四是要便于集成运动补偿处理。针对天线的连续运动问题,提出了一种改进的频率尺度变换(Modified Frequen-cy Scaling,MFS)算法,实现正侧视和小斜视角情况下快速的距离向处理。在非常适合去斜率信号实时成像的FS算法的基础上,结合调频连续波SAR特殊的信号模型,给出了严格的距离-多普勒域的信号解析式。分析了天线的连续运动导致的相位变化与方位向频率之间的关系。MFS算法通过在距离.多普勒域引入了一个新的相位因子校正了天线的连续运动。所引入的新的相位因子可以与原始FS算法中的块平移因子同时实现,因而该算法相对原始FS算法并不增加运算量。针对系统采样频率低的问题,提出了两种算法实现斜视角情况下的距离向处理。一种是MFS算法的扩展(Extension to MFS,EMFS)算法,这种方法从调频连续波SAR残余视频相位(Residual Video Phase,RVP)可以忽略的特点出发,在MFS算法的基础上通过增加斜置处理并修正相关相位因子来降低处理斜视角数据过程中引入的信号带宽。另一种是调频变换成像(Chirp Transform Imaging,CTI)算法,该算法同样从RVP可以忽略的特点出发,将距离徙动校正问题转化为一个非标准的Fourier变换问题。利用调频变换方法快速实现了这个非标准的Fourier变换。克服了MFS算法处理斜视角回波数据时存在的距离向频谱混叠问题。CTI算法效果与EMFS算法相当,但运算效率优于EMFS算法。针对扫频非线性的问题,提出了一种随多普勒中心变化的FS(Doppler CentroidDependent Frequency Scaling,DCDFS)算法实现存在扫频非线性时的斜视角情况下的快速距离向处理。分析了扫频非线性对距离向压缩性能的影响,给出了不同类型的扫频非线性情况下SAR成像对线性度的约束条件。在同时完成扫频非线性校正与距离徙动校正的过程中引入了一个与多普勒中心有关的因子,减小了斜视角的影响,从而降低了处理过程中引入信号的带宽,消除了距离向的频谱混叠。该算法可以直接推广到脉冲SAR的成像处理中。针对运动补偿对成像的制约问题,提出了一种随距离变化的步进变换(RangeVarying Step Transform,RVST)算法实现方位向数据的高效压缩。该算法利用典型的机载运动补偿通常在时域完成的特点,从时域出发,利用去斜率的方法实现方位向的压缩。由小型平台飞行高度较低的特点,考虑了斜视角随距离的变化关系,并在处理中补偿了这种变化。该算法同时分析了斜视角情况下调频斜率的非线性问题。给出了上述各种算法详细的推导过程及其解释,分析了各种算法的运算效率和处理误差,通过点目标仿真试验验证了各种算法的有效性。更多还原

【Abstract】 Frequency-modulated continuous-wave(FMCW) synthetic aperture radar(SAR) is a light-weight,cost-effective,high-resolution imaging radar,which is suitable for small platforms such as unmanned aerial vehicles.Compared with pulsed SAR,there are four problems in FMCW SAR real-time signal processing.Firstly,the continuous antenna motion will induce serious dilation in the received signal and the dilation can cause serious distortions in the reconstructed images.Secondly,dechirp-on-receive technique can be used to reduce the bandwidth of the received signal before analog-to-digital conversion,so the system sampling frequency is low.Thirdly,sweep frequency non-linearity in the transmitted signal deteriorates the range compression seriously.Fourthly,the motion compensation should be easily compatible with the image formation algorithm.A modified frequency scaling(MFS) algorithm is proposed to perform the range processing of side-looking or low squint FMCW SAR signal with the continuous antenna motion.Based on the frequency scaling algorithm,which is suitable for dechirped signal,the MFS algorithm gives the rigorous expression of range-Doppler domain signal.The relation between the phase resulting from the continuous motion and the azimuth frequency is analyzed.The MFS algorithm compensates the continuous antenna motion by introducing a new phase term in range-Doppler domain without extra computation load.Two methods are proposed to perform the range processing of squint FMCW SAR. The first is the extention to the MFS(EMFS) algorithm.According to the negligibility of the residual video phase(RVP) of FMCW SAR,the EMFS algorithm removes the range frequency aliasing by adding a skewing operation and modifying relative phase factors.The second is the chirp transform imaging(CTI) algorithm.Also based on the negligibility of the RVP,the algorithm transforms the problem of range cell migration (RCM) correction to the problem of a non-standard Foruier transformation.The CTI algorithm performs the non-standard Fourier trandformation by a chirp transform with low sampling frequency.The performances of the EMFS algorithm and those of the CTI algorithm are equivalent.The CTI algorithm is more efficency than the EMFS algorithm. A Doppler centroid dependent frequency scaling(DCDFS) algorithm is proposed to perform the range processing of squint FMCW SAR signal with non-linearity.The impact of transmitted signal non-linearity is analyzed.The linearity requirement for two different type phase en,ors in SAR imaging is given.The DCDFS algorithm introduces a factor dependent Doppler centroid to alleviate the impact of the squint angle when the nonlinearity and the RCM are compensated at the same time.At the position,the range frequency aliasing is removed.The algorithm can also be extended to pulsed SAR.A range varying step transform(RVST) algorithm is proposed to perform the azimuth processing of FMCW SAR signal.The algorithm performs the azimuth compressing by using the dechirping method from the characteristic of accomplishing the second-order motion compensation in azimuth time domain.Due to the large variation of the look angle in the small platforms’ FMCW SAR,the Doppler centroid can vary several hundred hetz from the near to the far range.The algorithm accommodates the variation of the Doppler centroid with the range distance by using a range-varying phase factor.The nonlinear chirp rate of the azimuth signal is also considered.The complete derivation of aforementioned algorithms is given.The computation load and processing error of the algorithms are analyzed.The algorithms performances are shown by point target simulation results.更多还原