【作者】 苏军海；

【导师】 邢孟道；

【作者基本信息】 西安电子科技大学 ， 信号与信息处理， 2009， 博士

【摘要】 雷达成像技术作为一种全天时、全天候、远距离的信息获取手段在国防以及民用领域均具有重大的应用价值,它大大提高了雷达信息获取能力,为目标识别创造了前所未有的机会,是雷达发展的一个重要里程碑。逆合成孔径雷达成像技术作为雷达成像技术的一个重要分支,主要应用于对空中或者空间以及舰船等目标的成像。成像的前提是检测到目标,及时、准确地发现目标也是非常重要的一环。本文主要研究宽窄带雷达的目标检测以及成像。论文内容可概括为如下五部分:第一部分,首先对目标跟踪所涉及的目标测速问题进行探讨。利用宽带包络对齐的包络移动量与窄带测距的距离之和得到目标的实际距离,并对其进行曲线拟合,从而得到目标的实际速度。相比于传统的窄带测速,利用宽窄带联合包络测速的方法得到的速度估计其估计方差更小,估计也更为精确。同时也考虑了将宽窄带包络和窄带信号的多普勒信息两者联合起来,借助于多普勒分析,可以达到更高的测速精度。第二部分,提出了两种窄带环境下的单目标检测方法。通过构造相应的相位补偿函数,实现信号的相干化,从而通过简单的快速傅立叶变换(FFT)来实现信号能量的积累。一种方法是利用稀疏Radon变换的方法进行参数的估计,而另一种方法则是通过修正的Keystone变换的方法来获得目标的运动参数。通过仿真实验可以看出,所提方法可以实现信号能量的有效积累,从而有利于目标的检测。第三部分,提出了两种适用于窄带环境下的多目标检测方法。基于信号分离的多目标检测方法和基于Clean技术的多目标检测方法,两者都是实现信号的相干化然后利用FFT检测目标,但是前者需要先对信号进行分离,涉及到调频率的搜索,因此其适用于SNR较高的情况,而后者由于Keystone变换对回波的SNR不敏感,更适用于SNR较低的情况。第四部分,对宽带信号环境下的单目标和多目标检测问题进行了系统的研究。首先利用宽带信号的相邻相关可以对回波信号的相位项进行降阶的性质,并结合时间-调频率变化以及二阶Keystone变换完成对信号的弯曲和走动的校正,从而最终实现对单目标信号能量的积累检测。对于多目标的情况,提出了两种方法对目标进行检测,第一种方法是对一个距离频域单元信号进行处理,第二种方法是对相关后的结果进行处理,两种方法都利用了Radon模糊变换,但是由于后者进行了相关处理,各个散射点的能量被集中了,因而更有利于目标的检测。第五部分,提出了一种基于多项式傅立叶变换并结合Clean技术的ISAR瞬时成像方法。该方法对机动目标的平动参数进行估计并进行补偿,然后对每一个距离单元的散射点参数利用多项式傅立叶变换进行搜索,并采用Clean技术以实现对多散射点参数的估计,接着利用获得的散射点参数可以得到任一时刻目标的距离多普勒像,并结合熵的概念提出一种选取最优成像时刻的方法。最后利用散射点的多普勒中心频率与调频率变化率之间的确定关系,提出一种估计ISAR转角的方法,从而实现ISAR像的定标。更多还原

【Abstract】 Due to its all-weather, all day/night and long range, radar imaging technique can enhance radar’s information acquisition capability greatly, thus exhibits great value in both civilian and military applications. It also provides strange change for target identification. At the same time, it is also a landmark for development of radar techniques. Inverse Synthetic Aperture Radar (ISAR) imaging technique is a main part of radar imaging technique. It may be applicated to image for air target, space target and ship target etc. Target detection is the principal problems for Inverse Synthetic Aperture Radar (ISAR) imaging. It is a very important step to detect target timely and exactly in ISAR applications. Target detection and imaging with wideband and narrowband radar are the primary content of this dissertation. It may be summarized as following five parts:1. The target velocity estimation for target tracking is firstly discussed. The real range of target from radar is acquired through the addition of the range of narrowband range gate (i.e. reference range for dechirp) and the envelope motion amount retrieving from the envelope alignment of wideband data, then the least square fit is applied to the real range, so that the velocity of the target is retrieved. Compared to the conventional narrowband velocity estimation method, the method which uses the wideband and narrowband data can obtain higher precision and smaller variation. At the same time, the envelopes of the wideband and the narrowband signal and the Doppler information of the narrowband signal are synthetized to obtain a higher precision velocity of the target.2. Two methods for the single target detection in the narrowband environment are proposed. By constructing the corresponding phase compensation function, the signal coherence can be achieved, and then the simple FFT operation can be taken to implement the target energy accumulation. The first method using Sparse Radon transform to obtain the target motion parameter, and the second method using the modified Keystone transform to achieve the target motion parameter. Through the simulation it can be seen that the methods can achieve the signal energy accumulation effectively and it is advantage for target detection.3. Two different methods are proposed for multi-target detection in the narrowband environment. The one is based on the signal separation and the other is based on the Clean technique. Both the methods can firstly achieve the signal coherence and then by FFT to detect target. But the former one need to separate the signal in the frequency domain, and the frequency modulation rate searching is required, so it may be applied to the high SNR condition. Whereas for the latter one due to the reason that the Keystone transform is not sensitive to SNR, so it may be applied to the lower SNR condition.4. The problem for target detection in the wideband environment is systematically studied. The characteristic of adjacent correlation of the wideband signal, i.e. decreasing order of the wideband echo signal, is applied, and then combining the time-chirp rate distribution and the second-order Keystone transform to achieve the correction of the range curve and range migration, and finally realize the signal energy accumulation detection for single target. Following for multi-target detection two methods are proposed. One is processed for a single range frequency cell; the other is processed for the result of adjacent correlation. Both the methods are adopted the Radon Ambiguity Transform (RAT), but due to the reason adjacent correction is applied in the latter one, the scatterer energy can be accumulated, so it is more advantage for target detection.5. A method which combines the Polynomial Fourier Transform (PFT) and Clean technique for Range-Instantaneous Doppler imaging Algorithm is proposed. The motion parameters of maneuvering target are firstly estimated and then they can be used for constructing a phase compensation function for target motion compensation, following for every range cell the PFT is applied to search the scatterer parameter, and then the Clean technique is used for multi-scatterer condition, so the RID imaging for every time can be obtained with the parameters estimated, and by introducing the definition of entropy a optimal time for RID imaging is presented. Moreover the relation of Doppler center frequency and the variation of frequency modulation rate is used for ISAR rotation angle estimation, and finally the scale for ISAR imaging can be achieved.更多还原