【作者】 陈柯；

【导师】 朱耀庭； 郭伟；

【作者基本信息】 华中科技大学 ， 通信与信息系统， 2010， 博士

【摘要】 作为一种新型干涉式阵列微波辐射成像系统,综合孔径辐射计采用稀疏的小口径天线阵列合成一个大的物理观测口径,降低了天线的体积与重量,且无需扫描即可实现对整个视场的瞬时成像,能够更好的满足实际应用的需求,具有很强的应用前景。然而越强的能力带来的限制也就越多。要实现良好的成像性能,综合孔径微波辐射计在实际工作中需要解决以下三个关键问题：①各种硬件非理想性因素引入的系统误差如何校正?②如何实现解决综合孔径亮温图像重建问题的高性能反演数值算法?③如何定量评估图像反演算法性能乃至系统整体成像性能?这三个问题的解决是层次递进的：综合孔径微波辐射成像系统的最终目的是得到高质量的亮温图像,因此对系统整体成像性能的判断,就需要依靠对亮温图像质量的定量评估来解决；而要从测量数据中重建高质量的亮温图像,就需要实现高性能的图像反演数值算法；而高性能的反演算法又需要准确的数据,这就必须以深入了解系统误差特性、实现对系统误差的校正为前提。本文针对综合孔径微波辐射成像系统以上三个关键问题的解决方法和关键理论开展研究,主要内容如下：首先本文以可见度函数为对象,根据系统误差作用于可见度函数的性质不同,将其分为系统加性误差、方位无关乘性误差和方位有关乘性误差三类,对前两者分别提出了利用外部参考场景和外部单辅助源的系统整体校正方法,并通过试验进行了验证。这种误差分析和校正思路的最大特点就是从系统整体出发,化繁为简,在不增加硬件复杂度的情况下实现对系统误差的整体校正。方位有关乘性误差导致综合孔径辐射成像系统的图像重建在数学上是一个病态的反问题,对其求解需要两大要素：①合适的图像反演数值算法；②获得系统响应G矩阵。在辐射计的低信噪比条件下,最常用的广义解会被噪声严重污染从而偏离所求解问题的真值。为解决这一问题,数学上提出了利用正则化参数在真值和噪声之间寻求平衡的正则化求解思想。本文在建立综合孔径系统图像重建问题的数学模型基础上,将数学上的正则化方法应用于综合孔径亮温图像重建问题的求解中,实现正则化的图像反演数值算法,再利用外部点源测量实际系统响应G矩阵,最后将两者结合,成功的应用于综合孔径亮温图像重建,实现了高质量的反演亮温图像。灵敏度是传统上衡量微波辐射计测量精度的主要指标,但并不适合对单帧亮温图像的质量进行评估。本文提出用图像信噪比作为图像评估指标,设定在均匀背景下存在单个目标的参考场景,图像信号强度定义为目标与背景的亮温差,图像噪声强度定义为均匀背景的波动,用背景的空间标准差来计算。这样定义的图像信噪比表征了单帧亮温图像的空间统计特性,因而能更好的实现对单帧图像的定量评估。本文将图像信噪比成功应用于综合孔径试验图像的评估中,对几种反演算法性能进行了比较,并反过来应用于正则化参数的后验判定,促进了正则化反演数值算法的研究。本文对综合孔径微波辐射成像系统及其关键理论和技术进行了较全面的描述,对其三个关键问题提出了自己的解决方法,并且对所有的理论和解决方法都给出了试验验证结果,这也是本文的特点之一。本文希望在综合孔径微波辐射成像系统的理论和实际应用之间搭起一座桥梁,促进该技术在各领域的广泛应用。更多还原

【Abstract】 As a new array interferometric microwave radiometric imaging systems, aperture synthesis radiometer (ASR) formed a large physical observation aperture by a thinned array of single small-aperture antennas, which reduced the antenna size and weight, and can instantaneously image for the entire field of view (FOV) without scanning, so it can better meet the needs of practical application and has a strong application prospects.However, more power associated with more restrictions. To achieve good imaging performance in practice, ASR need to address the following three key problems:①How to calibrate systematic errors introduced by various non-ideal factors of the hardware system?②How to realize image inversion numerical algorithm of high performance that solve the problem of aperture synthesis brightness temperature image reconstruction?③How to quantitative evaluation of the performance of image inversion algorithm as well as overall system imaging performance? These three problems are the progressive levels. The high-quality brightness temperature images is the final destination of aperture synthesis microwave radiometric imaging system, so it is necessary to realize quantitative evaluation of the inversion brightness temperature image for estimation of the performance of overall system imaging performance; and it is necessary to realize image inversion numerical algorithm of high performance for high-quality brightness temperature images; finally the image inversion algorithm requires the measured data of as accuracy as possible, which must be based on in-depth understanding of systematic errors characteristics and proper calibration of systematic errors. The research on the solutions and theory of three key problems of aperture synthesis microwave radiometric imaging system is the subject of this paper, the main contents are as follows:Firstly from the point of visibility function, in the paper the systematic errors were classified into three types by the roles in the visibility function:additive error、multiplicative errors of being independent of orientation、multiplicative errors of being dependent on orientation. The overall system calibration method for the first two errors is presented, which respectively made use of external reference scene and external single noise source, and has be verified by experiments. The most prominent feature of the error analysis and calibration is the idea of overall system that simplified the complicated problem, which achieve overall errors calibration without increasing system complexity.The inverse problem of aperture synthesis image reconstruction is ill-posed, which caused by the multiplicative errors of being dependent on orientation. Two major elements required for its solution are:①proper image inversion numerical algorithm;②system response G matrix. In the low SNR condition of radiometer, the usual generalized solution of the inverse problem may be completely corrupted by noise and therefore completely deprived of the exact solution. So the mathematical regularization methods were proposed to solve this problem, which made use of regularization parameter to achieve a balance between the noise and the true solution. In this paper, firstly the mathematical model of the inverse problem of aperture synthesis image reconstruction was establishment, secondly the mathematical regularization methods were applied to the solution of the problem of aperture synthesis image reconstruction and regularization image inversion numerical algorithm was realized, then system response G matrix was measured by using an external point source, finally the combination of both was successfully applied to ASR brightness temperature images reconstruction to get the high-quality inversion brightness temperature images.Traditionally, sensitivity is major performance parameter of microwave radiometer measurement precision, but it is not appropriate to quality assessment of single brightness temperature image. In the paper, the image signal to noise ratio (SNR) was present as a performance parameter of image assessment. A scene of single target against the uniform background was set the reference one, the difference of brightness temperature between the target and the background was defined as the signal intensity of image, the fluctuation in the background was defined as the noise intensity of image which was calculated by the standard deviation of the background. The image SNR defined a better performance parameter of quality assessment of single image based on its spatial statistical properties. In the paper, the image SNR was successfully applied to the assessment of the ASR images of imaging experiment, the performance of various image inversion algorithms were compared; contrariwise the assessment of images was also applied to the choice of regularization parameter and promoted the research of regularized inversion numerical algorithm.In a word, aperture synthesis microwave radiometric imaging system and its key technologies were comprehensively described, and the unique solutions of three key problems of ASR were present in the paper. One of the features of the paper is that experiment results were given for all the main theory and solution methods in the paper. The paper was expected to be a bridge between the theory and the practical application of aperture synthesis microwave radiometric imaging system, and can promote the wide application of the technology in various fields.更多还原