A Noisy SAR Image Fusion Method Based on NLM and GAN

The unavoidable noise often present in synthetic aperture radar (SAR) images, such as speckle noise, negatively impacts the subsequent processing of SAR images. Further, it is not easy to find an appropriate application for SAR images, given that the human visual system is sensitive to color and SAR images are gray. As a result, a noisy SAR image fusion method based on nonlocal matching and generative adversarial networks is presented in this paper. A nonlocal matching method is applied to processing source images into similar block groups in the pre-processing step. Then, adversarial networks are employed to generate a final noise-free fused SAR image block, where the generator aims to generate a noise-free SAR image block with color information, and the discriminator tries to increase the spatial resolution of the generated image block. This step ensures that the fused image block contains high resolution and color information at the same time. Finally, a fused image can be obtained by aggregating all the image blocks. By extensive comparative experiments on the SEN1–2 datasets and source images, it can be found that the proposed method not only has better fusion results but is also robust to image noise, indicating the superiority of the proposed noisy SAR image fusion method over the state-of-the-art methods.     

从SAR遥感图像中提取水域的一种双模式结合方法

水域是SAR图像的一个重要特征,针对SAR图像水域的检测问题,提出了一种利用形态学和信息跟踪双模式相结合的检测方法。采用形态学的开运算,提取基本的水域轮廓,再结合改进的信息编码,进行跟踪检测,进一步精确定位水域区域。实验结果表明,双模式结合的水域检测方法,可以较准确的提取SAR图像的水域区域,效果要优于单一模式的检测方法,而且跟信息熵方法相比较还提高了运算效率。     

合成孔径雷达图像的Contourlet域空间选择滤波降斑算法

合成孔径雷达(SAR)图像的相干斑噪声降低了影像分割和分类的效果,会导致影像判识困难。结合Con-tourlet系数的结构特点和SAR图像相干斑乘性噪声模型,提出了一种新的基于Contourlet变换的降斑算法。该算法利用信号与噪声在Contourlet域尺度间相关性的不同,通过空间选择滤波对系数进行分类,采用混合分布模型,实现了SAR图像的降斑处理。实验结果表明,该算法不仅能有效地抑制相干斑,而且还能有效地保护图像的细节特征。     

基于Contourlet变换的自适应SAR图像相干斑噪声抑制算法

为了保持高岛分辨率合成孔径雷达(SAR)图像中的纹理结构,提出了一种基于BivaShrink模型的Contourlet 域SAR图像相干斑噪声抑制算法.联合当前层和父层的Contourlet系数,通过计算局部方差一致性范数和区域能量比,自适应地确定方差估计区域的形状和大小.从而对原始图像方差进行最优估计.实验结果表明,算法在噪声的去除和结构信息等细节的保持上均不同程度的优于小波BivaShrink去噪算法和Contourlet 阈值去噪算法,主观效果和数值指标都有较好改进.     

一种新的基于纹理和空间分布特征的图像检索

提出一种新的基于纹理和空间分布特征的图像检索方法.将检索图像分块,采用平移和尺度不变小波对各图像子块进行分解,在改进的快速小波直方图算法基础上提取图像子块的小波直方图,并提取每个图像子块的小波信息熵和三阶中心距作为纹理特征.对小波信息熵和中心矩特征进行高斯归一化,并利用特征向量的欧氏距离计算图像的纹理和空间特征的相似度.基于纹理图像库和自然图像库的检索试验表明,该方法比基于快速小波直方图算法和对数极坐标变换检索算法具有较高的检索准确度.     

分形粗糙面合成孔径雷达成像研究

研究了分形粗糙面的成像问题. 分形粗糙面能够较好的逼近真实环境, 采用带限形式的Weierstrass-Mandelbrot函数建立了分形粗糙面几何模型, 对分形粗糙面参数的选取进行了讨论. 对大尺度粗糙面散射问题提出了一种基于大面元的Kirchhoff近似方法, 采用面元模型来计算粗糙面总的后向散射场以及每一个面元的后向散射场, 并对面元的尺寸选取进行了研究, 通过与解析解进行对比证明了该方法的正确性. 在分形理论建立的确定性粗糙面几何模型与面元的Kirchhoff方法获取的散射场的基础上, 采用正侧视条带式成像模式, 并选用距离多普勒算法对不同分形参数的粗糙面进行合成孔径雷达(SAR) 成像模拟, 结果显示从SAR像中可以清晰地观察到不同分形参数对粗糙面几何轮廓的影响. 该研究包括了从环境模型、电磁模型到SAR成像技术在内的完整的分形环境SAR像模拟过程, 仿真结果显示出分形环境的SAR像特点, 这对基于分形理论的自然环境的遥感探测以及参数反演具有一定的理论支撑作用.     

基于光谱特征的高光谱遥感影像检索

面向海量遥感信息管理中遥感影像检索的需求,以高光谱遥感信息为例对光谱特征应用进行了探讨,提出基于光谱特征的遥感影像检索包括基于点源示例的检索和基于面源示例的检索,检索中的关键问题是光谱特征提取与相似性度量。基于光谱曲线的相似性度量可以采用光谱角和光谱信息散度进行;基于光谱特征的检索可以通过提取反射和吸收光谱、光谱特征匹配的方法进行;而光谱曲线量化指标如中心矩、分维数和信息熵等的效果较差,不适宜应用于检索。     

A New Spectral Estimate for the SAR Image of the Ocean

We present a consistent theory of synthetic aperture radar (SAR) imaging of the ocean surface. Based on the composite model of microwave backscattering, the theory describes both the image itself and speckle noise, which is the factor preventing the interpretation of SAR imagery of the ocean. We propose a new spectral estimate, which, from the statistical viewpoint, eliminates completely speckle noise in the SAR image spectrum and does not deteriorate the SAR resolution. The new technique for estimating the spectrum is tested using the SAR data of the European satellites ERS-1 and ERS-2.     

Adaptive filtering for reduction of speckle in ultrasonic pulse-echo images

Current medical ultrasonic scanning instrumentation permits the display of fine image detail (speckle) which does not transfer useful information but degrades the apparent low contrast resolution in the image. An adaptive two-dimensional filter has been developed which uses local features of image texture to recognize and maximally low-pass filter those parts of the image which correspond to fully developed speckle, while substantially preserving information associated with resolved-object structure. A first implementation of the filter is described which uses the ratio of the local variance and the local mean as the speckle recognition feature. Preliminary results of applying this form of display processing to medical ultrasound images are very encouraging; it appears that the visual perception of features such as small discrete structures, subtle fluctuations in mean echo level and changes in image texture may be enhanced relative to that for unprocessed images.     

融合改进直方图PDE和二维Tsallis熵多阈值SAR分割

针对合成孔径雷达(SAR)图像中存在大量的相干斑噪声,对SAR图像进行分割易出现分割不精、边缘模糊等问题,融合改进的直方图PDE和二维Tsallis熵多阈值,提出了一种SAR图像分割算法。根据PDE直方图均衡化方法,将图像去噪与图像增强加权融合,利用各自权值调整去噪项与图像增强项;同时将二维Tsallis熵单阈值分割方法扩展到多阈值分割, 建立基于多阈值的选取方法,并引入萤火虫算法来求解最优阈值对,实现了二维Tsallis熵多阈值对去噪增强SAR图像的有效分割。仿真结果表明:与其他3种分割算法相比,该文算法在处理噪声大、灰度差值小的图像时具有较高的分割精度,PRI至少提升2.53%、VOI降低8.48%、GCE降低11.14%。     

高温下数字图像相关散斑最优成像探究

为了测量材料在高温甚至超高温下的力学性能,采用数字图像相关方法,并研究其在高温下的最优成像。采取不同的散斑制作方法,同时加入不同颜色的高温漆,在不同的温度节点,外加不同光源及相应的滤波片,采集并观察图像是否具有良好的对比度。普通的单色光源在800℃以后会逐渐失效,无法获取图像,而紫外光在1 200℃时依然可以获取较好的图像,且直接利用试件本身颜色作为底色效果更佳。采用紫外光照明可以实现DIC在高温环境下的测量。同时利用黑色或者蓝色散斑直接喷涂在试件上有着最佳的对比度,要优于常规的散斑制作方法。     

Determination of skeleton and sign map for phase obtaining from a single ESPI image

A robust method of determining the sign map and skeletons for ESPI images is introduced in this paper. ESPI images have high speckle noise which makes it difficult to obtain the fringe information, especially from a single image. To overcome the effects of high speckle noise, local directional computing windows are designed according to the fringe directions. Then by calculating the gradients from the filtered image in directional windows, sign map and good skeletons can be determined robustly. Based on the sign map, single image phase-extracting methods such as quadrature transform can be improved. And based on skeletons, fringe phases can be obtained directly by normalization methods. Experiments show that this new method is robust and effective for extracting phase from a single ESPI fringe image.     

新的Cauchy-Schwarz距离函数与多模态医学图像配准

信息论测度,特别是Shannon互信息是多模态图像配准的一种重要方法,但除了互信息之外,仍然存在其它的函数来实现这一任务。对互信息、Kullback-Leibler距离和Shannon不等式之间相互关系作了分析,根据这些关系和不等式理论,提出了新的Cauchy-Schwarz距离,并将这一距离测度用于多模态医学图像的配准处理。从计算速度、噪声容忍性、测度函数图形的特点和图像窗口大小影响等几个方面,通过MR和PET医学图像的实验分析,对新的Cauchy-Schwarz距离测度和典型的Shannon信息论测度进行了分析比较。实验结果表明,新的Cauchy-Schwarz距离测度函数同样可以用于多模态图像配准,而且有着更强的噪声容忍性和更为节省的计算量。     

扫频光学相干层析视网膜图像配准去噪算法

多帧叠加平均处理是去除扫频光学相干层析系统散斑噪声、获得较为清晰结构信息的有效方法,但眼睛的震颤、漂移、微眼跳等生理特性和系统光路特性会使图像之间存在错位,导致叠加效果不佳、结构稳定性差,为此本文提出一种基于灰度分布信息和目标几何信息相结合的配准算法。该方法根据图像平均灰度分布提取包含目标信息的感兴趣区域,通过相位相关算法和基于分段拟合的灰度投影算法的双重作用校正图像的平移变换;通过拟合视网膜上边界作为特征点迭代确定最佳旋转参数,并再次重新估计平移参数,实现图像的刚性配准;最后通过轴向扫描一对一映射法以能量函数为约束条件实现图像的非刚性配准。对活体兔眼进行实验,结果表明,本文算法配准后的叠加图像边界清晰,结构信息增强,信噪比和对比度平均有效提高一倍多。本算法适用于强噪声视网膜B-Scans图像的配准,能满足多种类型OCT系统的叠加成像需要,具有较高的鲁棒性和图像配准精度。     

基于自相关函数的非平面表面粗糙度的图像纹理研究

采用自相关函数对不同加工工艺形成的非平面工件表面粗糙度进行了研究。讨论了自相关函数及其扩展度参数与图像纹理特性的关系,构建了实验装置,利用图像处理软件对实验所得的激光散斑图像进行了处理,得到了自相关函数及其扩展测度参数随表面粗糙度的变化曲线。为研究非平面工件的粗糙度,提供了一种新的方法。     

利用激光散斑获取深度图

相对于普通灰度和彩色二维图像,深度图像可以得到物体的三维信息,使视觉识别和人机交互更加智能。国内外目前还没有低成本、公开的实时获取高质量深度图的方法。本文在对散斑图获取深度图原理研究的基础上,采取激光散斑的方式,运用块匹配的方法给出了一种大范围深度图的获取方法。首先,从原理上验证了块匹配方法的可行性;然后,分别从理论和实验两个方面对深度图的计算公式进行了推导和验证;再次,对深度图恢复过程进行了详细叙述,包括散斑图像的预处理和块匹配的过程;最后,给出了运用该块匹配方法得到的实验数据。实验结果表明,本文方法在物体距离相机50 cm左右时精度可以达到5 mm,200 cm时精度可以达到5 cm,可以满足室内大部分对象的识别要求。     

基于GoogLeNet和空间谱变换的高光谱图像超分辨率方法

为了提高高光谱图像的空间分辨率,提出了一种基于GoogLeNet和空间谱变换的高光谱图像超分辨率(SR)方法.设计出遥感图像的光谱SR框架,对图像中不同反射光谱进行提取;采用GoogLeNet的稀疏编码对粗像素光谱进行放大,并投影到高分辨率字典上,将潜在SR表示进行反转,以获得超分辨光谱;为了提高图像重构的保真度,利用...     

A Texture Combined Multispectral Magnetic Resonance Imaging Segmentation for Nasopharyngeal Carcinoma

Tumor segmentation from magnetic resonance imaging (MRI) is important for volume estimation and visualization of nasopharyngeal carcinoma (NPC). In some cases, segmentation using the general multispectral (GM) method often obtained poor results due to the high false positives caused by complex anatomic structures and serious overlap in feature space. In this study, a texture combined multispectral fuzzy clustering (TCMFC) segmentation algorithm was proposed. A texture measure of T1-weighted (T1) MR image was introduced by calculating the two-order central statistical information of every pixel within a window after the window convolution operation. The texture measure and the intensities in T1 and contrast-enhanced T1 images formed the new 3-D feature vector for fuzzy clustering implemented by semi-supervised fuzzy c-means (SFCM). Testing showed that by reducing the false positives significantly, the TCMFC method achieved improved segmentation results, compared with the GM method.     

Digital image correlation with gray gradient constraints: Application to spatially variant speckle images

As a carrier of local deformation information, speckle pattern inside a subset is usually crucial for surface displacement acquisition based upon a digital image correlation (DIC) method, since both accuracy and precision of DIC method are closely related to the amount of speckle information in a subset. Although some comprehensive theoretical frameworks have been developed to estimate the quality of local speckle patterns, it is still a great challenge how to effectively integrate the subset speckle information into the well-developed correlation criteria used for DIC. By means of a well-designed square window function, we here propose the concept of continuous subset in order to modulate subset size in a continuously derivable manner. Afterwards, we further develop a new constrained zero-normalized sum-of-squared differences (CZNSSD) criterion and construct the corresponding iterative algorithm, based on which the subset size involved can be automatically determined according to the necessary amount of speckle information. Numerical results of synthetic speckle images indicate that the set of algorithm can enhance the accuracy and precision of displacement measurement, especially for spatially variant speckle images.     

局部方差在图像质量评价中的应用

将灰度图像的局部方差分布(QLS)作为表征图像结构信息的一个重要特征,对局部方差分布矩阵进行奇异值分解,计算得到相应的奇异值特征向量;通过计算降质图像与原参考图像局部方差矩阵奇异值特征向量的夹角大小度量两图像的结构相似度,实现了对降质图像的质量评价。实验结果表明:局部方差分布更能突出图像的结构特征,评价结果优于传统的均方误差(MSE)、峰值信噪比(PSNR)、结构相似度(SSIM)以及直接评价图像像素分布的奇异值分析(SVD)等方法,与人眼视觉感知效果的一致性较好。