Optimum Windows for Image Registration

Two-dimensional cross-correlation techniques are applied to the problem of image registration under the assumption of small geometric distortion. Optimum window functions are derived for two performance measures of interest: peak-to-sidelobe ratio and mean-square registration error. The latter is examined in terms of the contribution caused by noise and the contribution caused by geometric distortion. A generalized Lagrange multiplier approach is used to derive approximate solutions assuming random images. The case of Gaussian autocorrelation functions is examined in detail. Results of applying the theoretically derived window functions to real data are presented, showing significant improvement in correlator performance.     

Image Correlation with Geometric Distortion Part II: Effect on Local Accuracy

The effect of geometric distortion on the local accuracy of the image registration algorithms using cross correlation is presented. Using a probabilistic model describing images as homogeneous random patterns, expressions for the mean and covariance of the local error vector in terms of image and noise autocorrelation functions, geometric distortion, and reference image area are derived. The geometric distortions considered are those represented by an affine transformation of image coordinates. It is shown that for a fixed geometric distortion there is an image size (integration area) that minimizes the local error. The optimum area decreases with increasing geometric distortion.     

Image Correlation with Geometric Distortion Part 1: Acquisition Performance

The degradation in acquisition probability that occurs when cross correlation is used to determine the offset of two images of the same scene that differ by a relative geometric distortion is presented. The geometric distortions considered can be represented by a general affine transformation of image coordinates. The analysis shows that for a given geometric distortion there is an image size and shape that minimizes the probability of false acquisition. The results are derived for images modeled as random patterns with arbitrary auto-correlation functions. The results are illustrated for images with Gaussian autocorrelation functions.     

复杂地形下基于三视约束的景象匹配方法

 景象匹配通过将飞行器实拍图像与机载的基准图像比较而进行导航。当飞行高度较低时,由地面大起伏引起的实时图几何变形会严重影响匹配性能,造成匹配精度下降甚至出现误匹配,这一问题严重制约了现有景象匹配系统的应用范围,比如无法应用于建筑林立的城镇地区和梯田、丘陵地带。本文首先推导了小视场内无几何变形的高空竖直透视投影模型;在此基础上,依据三视约束关系,把相邻两幅下视实时图像的特征赋给该透视模型所成的像,得到消除几何变形的特征模板,再和基准图进行匹配。因此,所提方法不仅能处理由飞行器姿态引起的实时图几何变形问题,而且适用于城镇等地面存在较大起伏的地形。理论分析和仿真实验表明,该算法取得了较好的结果,具有良好的环境适应能力,实用性较强。     

Real-time mosaicking for infrared videos from an oblique sweeping camera

Image mosaicking is widely used in Geographic Information Systems (GISs) for large-scale ground surface analysis. However, most existing mosaicking methods can only be used in offline processing due to the enormous amounts of computation. In this paper, we propose a novel and practical algorithm for real-time infrared video mosaicking. To achieve this, a novel fast template matching algorithm based on Sum of Cosine Differences (SCD) is proposed to coarsely match the sequential images. The high speed of the proposed template matching algorithm is obtained by computing correlation with Fast Fourier Transform (FFT). We also propose a novel fast Least Squares Matching (LSM) algorithm for inter-frame fine registration, which can significantly reduce the computation without degrading the matching accuracy. In addition, the proposed fast LSM can effectively adapt for noise degradation and geometric distortion. Based on the proposed fast template matching algorithm and fine registration algorithm, we develop a practical real-time mosaicking approach which can produce seamless mosaic image highly efficiently. Experiments on synthetic and real-world datasets demonstrate that the proposed algorithm is not just computationally efficient but also robust against various noise distortions.     

Distortion Generated in Angle-Modulation Systems by Phased Arrays

The case of linear, uniformly weighted phased arrays is examined via time and frequency domain analyses. Bounds that must be placed on array length, modulation frequency, modulation index ?, and scan angle relationships if excessive distortion is to be avoided in wideband angle-modulated communication systems are established. Distortion is shown to consist of odd harmonic terms. It is also shown that, for one class of equivalent RF signals, phased arrays produce approximately three times as much distortion in frequency-modulation (FM) systems as in phase-modulation (PM) systems. Graphs of distortion plotted as functions of signal and array parameters show that, for practical array sizes, distortion is a monotonically increasing function of the product of L and sin ? where L is the length of the array expressed in modulation wavelengths ?m and ? is the scan angle. In PM systems, distortion also increases monotonically with the modulation index ?. Plots of distortion versus L sin ? show that even relatively small arrays can produce intolerable distortion levels in wideband systems; e.g., an FM system having ? = 3,L = 0.35?m, and ? = 60 degrees exhibits approximately 20 percent distortion.     

图像配准关键技术综述

图像采集设备的快速发展带来了图像种类和数量的迅速增加,图像拼接、融合等视觉系统提高了对图像配准的技术需求。图像配准利用原始图像数据或图像特征信息对输入图像的重叠区域进行变换和对齐,是为后续图像拼接、图像融合等图像处理任务提供操作基准的核心技术,是当前计算机视觉领域的热点研究问题。首先介绍了图像配准技术的主要应用领域及重大发展前景,然后说明了图像配准技术的基本框架和技术难点,详细阐述了基于特征以及基于区域的图像配准技术的实现算法及关键技术,分析了不同配准算法的优缺点及适用性,并对图像配准技术的发展前景进行了讨论和展望。     

不同飞行状态下施翼与大气湍流的干扰噪声

本文简述了一种直升机旋翼与大气湍流入流相干扰产生的噪声预估方法。这一方法考虑了桨叶上非定常载荷的弦向及展向非紧致性，桨叶－桨叶载荷之间的相关，通过结合大气湍流能量谱模型与快速畸变湍流收缩模型来确定旋翼平面处非均匀、各向异性湍流场的特性，使得这种方法可用于直升机在真实飞行状态下的旋翼与湍流场干扰噪声计算。利用文献中模型旋翼试验的结果，验证了本方法的可行性，并计算了某型直升机在六种飞行状态下的流场畸变情况及旋翼与湍流干扰的远场噪声谱，结果表明：在准悬停及低速爬升状态，湍流场的畸变较大，而在中、快速前飞状态，湍流场的畸变最小，在低速爬升状态，线性总声压级最小。     

基于路面分割的高精度地图创建优化方法研究

高精度地图主要利用已采集图像的地面信息生成,但是在真实环境中图像的地面信息容易受动态障碍物遮挡,同时GPS难免会有抖动误差,导致地图拼接效果并不理想。针对上述问题提出了一种基于路面分割的动态障碍物去除与图像配准方法。使用深度学习对全景图像进行语义分割并提取路面信息,在去除动态障碍物干扰后,利用路面特征进行图像配准。融合GPS与里程计信息进行定位优化,利用多帧图像叠加填补空缺形成地图。最终,实验验证了该方法在去除动态障碍物的同时也提高了地图的精度。     

进气总压总温畸变对压气机稳定性的影响

本文对 Melick提出的进口总压周向稳态畸变对压气机稳定性裕度损失的估算方法作了改进 ,并加以推广应用于进口总温周向稳态畸变的情况。在此基础上 ,本文提出了总温和总压组合畸变下压气机稳定性裕度损失的预测方法。以某型双轴涡喷发动机为例作了某些具体的计算。     

Algorithms for Fast Image Registration

The automatic determination of local similarity between two images (image registration) is one of the most fundamental problems of image processing and pattern recognition. A class of registration algorithms that are reasonably efficient and robust for translational displacement has been considered to determine relative shift between reference and search images. Stochastic image models defined on a rectangular region of support are used to determine feature vectors associated with reference and search images. A new measure, namely, coefficient of variation, is defined to take into account effects of contrast and sharpness of the images. Based upon this measure, a computationally efficient two-stage algorithm is obtained by combining the image-model based algorithm with a template matching technique. Simulation results with several synthetic and real images are presented to evaluate the performance of the algorithms.     

Image Registration: Similarity Measure and Preprocessing Method Comparisons

An experimental comparison of several similarity measures and preprocessing techniques used for the registration of temporally differing images is carried out. It is found that preprocessing of the images via a gradient operator improves the registration performance. This is in agreement with a derived optimal processor (described in the Appendix) based upon image and temporal difference characteristics.     

遥感图像的亚像素配准技术研究

高分辨遥感技术在军民领域应用广泛,图像配准技术是高分辨遥感图像的预处理技术。实际的机载和星载遥感成像仪,由于自身成像原理和成像仪平台抖动原因,出现了遥感图像亚像素偏移误差。对亚像素图像配准技术进行系统研究,对高精度亚像素配准算法进行研究分析,并对其算法进行仿真、处理,对仿真的结果进行系统分析,通过误差分析提出配准误差修正算法,修正算法对亚像素配准进行了很好的改进,达到遥感图像预处理的目的。     

航空发动机进气道出口截面喘振波的分布

用扰流插板使涡扇发动机进气道内产生不对称分布的总压畸变并使发动机出现喘振。在进气道出口即发动机进口截面测取两种不同转速下发动机的喘振压力波。应用小波分析奇异点理论,采用模极大值原理和高阶Daubechies小波多尺度分析,对喘振压力信号进行处理,得到发动机喘振波在其进口截面的波形分布特征。揭示了在两种转速下,喘振波在截面上的分布相似,其频率和幅值与转速成正比。截面中心部分表现为正脉冲,周围部分为负脉冲。在垂直方向上的幅值较大,而两翼较小。对该分布特征进行了分析,初步的推论为:高压压气机引起喘振,稳态压力畸变最严重的区域和未受到畸变扰动的区域,喘振波最为强烈。     

Super-resolution reconstruction of astronomical images using time-scale adaptive normalized convolution

In this work, we describe a new multiframe Super-Resolution (SR) framework based on time-scale adaptive Normalized Convolution (NC), and apply it to astronomical images. The method mainly uses the conceptual basis of NC where each neighborhood of a signal is expressed in terms of the corresponding subspace expanded by the chosen polynomial basis function. Instead of the conventional NC, the introduced spatially adaptive filtering kernel is utilized as the applicability function of shape-adaptive NC, which fits the local image structure information including shape and orientation. This makes it possible to obtain image patches with the same modality, which are collected for polynomial expansion to maximize the signal-to-noise ratio and suppress aliasing artifacts across lines and edges. The robust signal certainty takes the confidence value at each point into account before a local polynomial expansion to minimize the influence of outliers. Finally, the temporal scale applicability is considered to omit accurate motion estimation since it is easy to result in annoying registration errors in real astronomical applications. Excellent SR reconstruction capability of the time-scale adaptive NC is demonstrated through fundamental experiments on both synthetic images and real astronomical images when compared with other SR reconstruction methods.     

An approach for machining distortion measurements and evaluation of thin-walled blades with small datum

Inspection techniques for aero-engine blades are a hot topic in industry. Since these blades have a sculptured surface and a small datum, measurement results may deviate from an actual position. There are few proper approaches compensating for non-uniform distribution errors that are within specified tolerance ranges. This study aimed to develop a meshing structure measur-ing approach for the distortion of blades via non-contact optical 3D scanning. A rough measure-ment and a registration procedure are initially adopted to rectify the coordinate system of a blade, which avoids the initial coordinate system errors caused by the small datum. A measurement path with meshing structure is then unfolded on the blade surface. For non-uniform distribution errors, the meshing structure measurement is more visual and clear than the traditional constant height curves method. All measuring points take only 7 min to complete, and the distribution of error is directly and accurately presented by the meshing structure. This study provides a basis for future research on distortion control and error compensation.     

Signal Sequence Detection Given Noisy, Common Background Image Sets

The optimum processing (likelihood functional) is found for a set of M images {Zm = Sm + Y + Nm}, each the sum of a member Sm of a signal sequence {Sm}, due to an object to be detected and its parameters estimated, a sample function Nm of a noise field {Nm}, and a sample function Y of a common background field {Y}. The noise fields {{Nm}} are independent, zero mean, white Gaussian fields, all independent of the background field {Y}; the latter is assumed to be either 1} completely unknown or of known mean and covariance functions with 2) a certain fluctuation property or 3) Gaussian. Three equivalent forms of the optimum processing are found: 1) a summation of generalized matched filterings of the images, 2) a summation of matched filtering of certain generalized differences of the images, 3) a summation of ?estimator-correlator? type filterings. The detection performance and optimum signal/image selection under the Neyman-Pearson criterion is given and the singularity of the ({{Nm = O}} and M > 1) case noted. It is shown that optimum processor and signal design can completely eliminate any effect of the background on detectability (M > 1). The Cramer-Rao lower bound for the signal parameter estimates meansquare error is given along with an example; optimum signal/image selection in the single parameter case is discussed.     

基于不变特征和映射抑制的航拍视频图像配准

针对航拍视频图像在光学变换和几何变换下特征匹配容易产生误匹配,以及匹配精度不一致等问题,提出一种结合不变特征和映射抑制的航拍视频配准算法。首先,利用Harris检测器在尺度变化下的局部区域内选择出最稳定角点,并利用分区检测方法提取出分布均匀的角点作为待配准点。然后,计算以特征点为中心邻域窗口的描述符,通过比较两幅待配准图像特征点邻域描述符的欧式距离,提取出初始匹配点对。对匹配对点集建立D-三角网图,利用映射抑制方法提取出精确度最高的少数匹配点,并对模型参数进行估计,完成视频帧间的配准。大量实验表明,所提方法对航拍视频图像配准具有较强的适应性且配准精度高。     

基于平面模板自由拍摄的双目立体测量系统的现场标定

 摄像机标定是基于光学摄像的立体测量技术中的一个十分重要的步骤，标定精度直接影响系统测量的精度和稳定性。本文提出并实现了一种简便易行的双目立体测量系统现场标定方法：利用一块特殊设计的具有不同大小圆形特征点的平面标定板，在无需控制任何运动参数的情况下，双目测量系统只需对标定板在不同角度自由拍摄一组图像即可方便地实现系统标定。该方法在已有单摄像机标定算法的基础上，加入对双摄像机相对位置和姿态的优化，同时考虑了镜头的非线性畸变，达到了较好的标定结果。另外，提出的图像点与其空间点的对应算法具有良好的稳定性，即使由于环境或摄像机摆放位置等因素的影响，一些标定板上的特征点不能被摄像机拍摄到，或者不能被正确识别时，仍然能进行标定工作。标定实验和系统标定后的三维重建结果验证了该方法的有效性。     

A Note on Dynamic Range of Quantizers

Based on the relationship between the inputand output correlation functions for a symmetric2N step optimum quantizer withGaussian noise inputs a computer programwas generated to obtain plots of the inputnumber of levels versus dynamic range of thequantizer. For it to be physically useful thedynamic range should be related to the distortionin the fundamental band and thedesired harmonic suppression. A definitionbased upon these considerations is suggestedand curves are plotted for two different fundamentaldistortion levels and three differentlevels of harmonic suppression. It is foundthat, for low levels of harmonic suppression(0-20 dB range), the dynamic range can beincreased by tolerating a higher level of fundamentaldistortion. In the medium range ofharmonic suppression (20-35 dB), any increasein the dynamic range due to higherfundamental distortion levels disappearsfor high level quantizers (typically 4 bitsor more). For still higher harmonic suppression(40 dB or higher), the dynamic range ofthe quantizer is independent of the acceptablefundamental distortion.