基于压缩感知的多光谱图像去马赛克算法

针对目前多光谱图像去马赛克算法存在计算量大、效率低的缺点,本文提出一种基于压缩感知的多光谱图像去马赛克算法。首先,分析去马赛克与压缩感知问题的等价性,建立基于压缩感知的去马赛克模型;然后,采用离散余弦变换构建压缩感知的稀疏基,将去马赛克问题转化为压缩感知的信号重构问题;最后,采用改进的光滑0范数和修正牛顿法的重构算法求解去马赛克问题,得到重构的多光谱图像。仿真实验表明,相对于基于克罗内克压缩感知和组稀疏两种算法,本文算法提高了重构的多光谱图像的峰值信噪比,能有效减少对比算法重构多光谱图像中出现的锯齿现象,改善了重构图像具有更好的视觉效果。实验结果验证了本文算法的有效性。     

基于压缩感知的高光谱遥感影像重构方法研究

高光谱遥感影像包含丰富的空间、辐射以及光谱信息,同时海量的数据也引发了高光谱成像技术在传输和存储方面的诸多问题。针对这一问题,根据高光谱遥感影像谱间相关性强的特性,提出了一种结合谱间多向预测的基于压缩感知的高光谱遥感影像重构方法。首先,根据高光谱遥感影像的谱间相关性对高光谱遥感影像的波段进行分组,每组确定一个参考波段,使用平滑l_0范数算法重构每组的参考波段。其次,根据重构恢复的相邻组内的参考波段,建立了一个非参考波段预测模型,用来计算非参考波段的预测测量值;然后,计算实际测量值与预测测量值的差值,使用SL0算法重构该差值得到差值向量;最后,利用得到的差值向量迭代更新预测测量值,直到恢复该波段原始图像。仿真实验结果表明,该方法提高了高光谱遥感影像的重构效果。     

基于小波域的压缩感知超分辨率重建方法

为了提高超分辨率图像的重构效果,提出了一种基于小波域的压缩感知超分辨率重建方法.首先采用小波变换对低分辨率图像进行分解,得到低分辨率图像的低频子带与高频子带,然后采用压缩感知技术分别对低频子带和高频子带进行重建,并通过小波逆变换得到最终重建图像,最后在MATLAB 2012平台上进行仿真实验.结果表明,相对于其他图像重建方法,本文方法的重建图像视觉效果更加理想,不仅提高了低分辨率图像重建精度,而且提高了图像重建效率,可以满足图像处理的实时性、在线性要求.     

Compression method for non-stationary signals based on compressive sensing

A new approach to compress non-stationary signals is proposed in this paper. The sparse basis of non-stationary signals is constructed at first and then compressive sensing technique is used to decrease enormously the number of samples in the process. The reconstructed signal can well approximate the original signal in time domain, frequency domain as well as the time-frequency domain. Computer simulation on linear frequency modulated (LFM) signal shows the validity of this novel method.     

THz隐身目标无源遥感建模及仿真分析

隐身目标由于其雷达目标反射截面太小,仅靠主动探测手段,其有效性受到限制,主被动联合探测是一个有效尝试。本文就被动探测方向的特征问题进行论证分析,采用“am”大气模型,就地基、空基隐身飞机目标的辐射探测任务要求进行THz无源遥感建模和仿真,并结合遥感系统进行综合指标论证,给出实现其任务目标的基本要求(包括频点、天线口面以及平台要求)。     

基于光滑0范数压缩感知的多光谱图像去马赛克算法

提出了一种基于压缩感知(CS)的多光谱滤波阵列(MSFA)的多光谱图像去马赛克算法(DMA)。 首先,通过将MSFA采样得 到马赛克图像的过程等效为CS理论中的感知矩阵采样的过程,并充分利用多光谱图 像的空间和谱间 相关性,通过在三维空间傅里叶基上对多光谱图像进行稀疏表示;然后由随机MSFA模式和CS 理论构造的测量矩阵对多光谱图像进行观测投影,最后采用CS重构算法求解0范 数下的最优化问 题,从而得到多光谱图像的稀疏表示系数。给出对算法性能的评估数据和Matlab仿真 图片。实验结果证明,本文算法的峰值信噪比(PSNR)值高于克罗内克CS(KCS)和组稀疏(GS)两种算法,且有效地减少了上述两种算法中出现的模糊现 象,改善了图像的视觉效果。     

基于压缩感知的正六边形CFA模式彩色图像去马赛克方法

针对基于四边形排列的去马赛克(Demosaicking)的 传统方法存在拉链现象和虚假色等问题,本文尝试将更加符合人眼视觉特性的六边形采 样方式应用于彩色图像成像 系统,并从图像稀疏特性角度出发,提出基于压缩感知(Compressive sensing,CS)框架的 彩色图像去马赛克方法。本文方法 充分挖掘了彩色分量间和分量内的稀疏特性,可使复原图像的纹理细节与色彩更加逼真,有 效地避免了拉链现象和虚假色现象。实验结果验证了本文方法的有效性。     

Terahertz Transmission Imaging with 2.52 THz Continuous Wave

In this article, two terahertz transmission imaging systems are built with a 2.52 THz continuous wave laser and two types of sensors. One is array scanning system using a 124×124 pyro-electric array camera as the detector; the other is a point-wise scanning system utilizing a Golay cell as the detector. The imaging speed and quality is briefly analyzed. Terahertz （THz） imaging results demonstrate that the array scanning system has higher imaging speed with lower resolution. The point-wise scanning system has higher imaging quality with lower speed.     

太赫兹波传播过程中横模分布的成像

研究了焦点位置附近太赫兹波的横模分布,目的在于了解系统中太赫兹波的横向分布情况,为进行太赫兹光谱和成像实验提供依据。通过太赫兹波逐点扫描成像,对经常被用来放置测试样品位置附近的太赫兹波横模分布情况进行了定性的分析。利用“狭缝法"测量了太赫兹光束的束宽。结果表明,太赫兹波在焦点附近不同位置的横向分布有较大差异,在焦点处对应于最长波长1.5 mm的最小光斑直径约为1.0 mm,这说明太赫兹波的聚焦有一定的限度。实验确定了系统中放置样品的最佳位置,对开展太赫兹光谱和成像研究有参考价值。     

一种快速的压缩感知信号重构算法

重构算法是压缩感知技术的重要环节之一,文中针对现有重构算法收敛速度较慢的问题,提出了一种适用于压缩感知的快速重构算法。该方法的思想是在求解过程中,设计一种有效的步长迭代方案,以此来更新由梯度Lipschitz指数确定的迭代步长,再利用更新后的步长对原始信号的稀疏域表示向量进行迭代收缩,提高收敛速度。实验结果表明,相比传统的正交匹配追踪(OMP)算法、固定步长的l1范数重构算法,该方法在保证信号恢复精度的前提下,具有更快的收敛速度和更高的重构精度。     

改进的基于拉普拉斯先验的贝叶斯压缩感知算法

贝叶斯压缩感知(Bayesian Compressed Sensing,BCS)通过稀疏贝叶斯回归模型中相关向量机(Relevance Vector Machine,RVM)的学习来解决压缩感知(Compressed Sensing,CS)中的信号重构问题。本文通过修正基于拉普拉斯先验BCS的噪声模型,较好地实现了含噪CS信号的重构。它主要利用稳健型相关向量机(Robust RVM,RRVM),改进了基于拉普拉斯先验的BCS算法。它通过对每个观测噪声方差系数进行最优化估计,来消除内外部噪声对信号重构的影响。相关的仿真验证了在外部脉冲噪声以及内部高斯白噪声共同干扰条件下,相比原始BCS算法,改进算法具有更好的重构性能和稳定性。     

近程目标毫米波全息成像算法及仿真

为解决近程毫米波雷达目标探测过程中产生的回波相位干涉问题,引入全息技术研究近程目标的毫米波散射特性成像.该方法利用相位补偿因子,在波数域补偿电磁波的波前球面弯曲,反演目标的真实几何形状特征.给出了单目标散射特性一维成像的仿真实验,验证全息成像算法的有效性.对有无相位补偿情况下的多目标散射特性一维成像进行比较,结果表明该...     

毫米波安检成像雷达设计

针对实际应用的安检技术的局限性,本文介绍一种基于FMCW的体制的安检成像雷达设计。雷达采用合成孔径成像算法对待检人员身上的隐匿目标进行探测,使用二维、三维仿真目标和Radarsat—I雷达数据对雷达系统和成像算法进行检验,计算结果证明了系统设计和成像方法的有效性。     

一种毫米波全息成像系统方案及其仿真

提出了一种双频毫米波全息成像系统方案。该方案不仅可以准确地测定目标与成像雷达天线口径的距离,而且可以对一定距离但事先未知的目标进行全息成像。相比于需要放置在预知位置的美国太平洋西北国家实验室研究的单频和宽带毫米波全息成像系统,该方案具有更为广泛应用的潜力,不仅可以用于个人携带的藏匿武器或者违禁物品的检查,还可以作为毫米波成像雷达用于军事目的。文中根据该成像系统确定了毫米波全息成像方法,建立了相应算法,给出了成像仿真结果。该方案可对实际系统设计提供一定参考。     

Saliency-guided compressive sensing approach to efficient laser range measurement

The acquisition of laser range measurements can be a time consuming process for situations where high spatial resolution is required. As such, optimizing the acquisition mechanism is of high importance for many range measurement applications. Acquiring such data through a dynamically small subset of measurement locations can address this problem. In such a case, the measured information can be regarded as incomplete, which necessitates the application of special reconstruction tools to recover the original data set. The reconstruction can be performed based on the concept of sparse signal representation. Recovering signals and images from their sub-Nyquist measurements forms the core idea of compressive sensing (CS). A new saliency-guided CS-based algorithm for improving the reconstruction of range image from sparse laser range measurements has been developed. This system samples the object of interest through an optimized probability density function derived based on saliency rather than a uniform random distribution. Particularly, we demonstrate a saliency-guided sampling method for simultaneously sensing and coding range image, which requires less than half the samples needed by conventional CS while maintaining the same reconstruction performance, or alternatively reconstruct range image using the same number of samples as conventional CS with a 16 dB improvement in signal-to-noise ratio. For example, to achieve a reconstruction SNR of 30 dB, the saliency-guided approach required 30% of the samples in comparison to the standard CS approach that required 90% of the samples in order to achieve similar performance.     

Fiber-Drawn Metamaterial for THz Waveguiding and Imaging

In this paper, we review the work of our group in fabricating metamaterials for terahertz (THz) applications by fiber drawing. We discuss the fabrication technique and the structures that can be obtained before focusing on two particular applications of terahertz metamaterials, i.e., waveguiding and sub-diffraction imaging. We show the experimental demonstration of THz radiation guidance through hollow core waveguides with metamaterial cladding, where substantial improvements were realized compared to conventional hollow core waveguides, such as reduction of size, greater flexibility, increased single-mode operating regime, and guiding due to magnetic and electric resonances. We also report recent and new experimental work on near- and far-field THz imaging using wire array metamaterials that are capable of resolving features as small as λ/28.     

Channel estimation method based on compressive sensing for FBMC/OQAM system

In mobile-to-mobile sensor networks,the channel estimation for FBMC/OQAM system can be investigated as a compressive sensing problem to raise frequency spectrum efficiency by exploiting the sparse nature of wireless channels.Firstly,a novel orthogonal matching pursuit algorithm with selection weak strategy and regularization based on Tanimoto coefficient (T-SWROMP) was proposed to improve the accuracy of LS channel estimation.Then,T-SWROMP methods with auxiliary pilot and coding were used to estimate channel frequency response for FBMC/OQAM system.The experimental results demonstrate the proposed method has lower complexity than the traditional SWOMP method.In addition,it achieve best performance among the traditional OMP,SWOMP and ROMP methods under dual-selective channels.     

基于改进SVA和压缩感知的SAL旁瓣抑制算法

针对机载合成孔径激光雷达实测数据成像中旁瓣较高的问题,提出一种新旁瓣抑制算法。压缩感知理论表明,稀疏信号恢复重构过程的同时,信号旁瓣会被压低,但合成孔径激光雷达图像不是稀疏的。针对这一问题,提出了一种基于改进SVA（Spatially Variant Apodization）和压缩感知重构SAL图像的旁瓣抑制算法。首先,利用改进SVA算法将SAL图像变稀疏,然后再利用压缩感知算法对稀疏图像进行恢复。分别对SAL仿真数据和实际高旁瓣SAL复图像进行抑制旁瓣处理。仿真结果表明:该算法能够在保证主瓣不被展宽的前提下有效抑制SAL旁瓣。     

Distribution agnostic Bayesian compressive sensing with incremental support estimation

In most compressive sensing algorithms, such as L1-optimization and greedy family techniques, the only a priori information utilized in the reconstruction procedure is the sparsity information. Meanwhile, there exists another family of techniques based on the Bayesian strategy, which considers comprehensive a priori statistical knowledge of the sparse data. This feature resulted in more increased attention to this category of algorithms. One member of the Bayesian-based family of compressive sensing reconstruction algorithms is the support Agnostic Bayesian Matching Pursuit, which is agnostic to support distribution.However, its high computation complexity in determining the set of dominant supports makes this algorithm unfeasible for practical applications such as wireless sensor networks (WSNs). Due to the special conditions of WSNs, consists of limited-power sensors, developed algorithms for them must have the least possible amount of computations. Given this, in this paper, we propose a Bayesian-based method with incremental support detection for distributed sparse signal recovery, which considerably reduces computational complexity. In the proposed method, in a network of sensors, sparse signal reconstruction from noisy measurements is done distributively and in the form of incremental cooperation. So, the number of required computations will be significantly reduced, which will result in a fast approach. The computer simulations show the superior performance of the proposed incremental Bayesian recovery method.

     

基于二维压缩感知的定向遥感和变化检测

基于高斯测量矩阵的一维压缩感知测量数据不仅能很好地保持稀疏信号的能量信息, 也能够很好地继承稀疏信号的方向信息.但是在一维压缩感知模型中方向信息无法应用于稀疏信号的重构和检验.针对遥感影像中变化区域稀疏的特点提出了二维压缩感知模型.并利用能量和方向信息构建了基于二维压缩感知的稀疏信号重构算法(2DOMP).理论分析和实验结果证明, 2DOMP算法的信号重构能力更强.同时根据压缩感知恢复稀疏信号只需要很少测量数据的特性提出了定向遥感和定向变化检测的概念.