用雅可比-傅里叶矩进行图像描述

提出了雅可比—傅里叶矩,它是用雅可比多项式作为径向函数,用傅里叶因子作为角向函数构造而成的,是广义正交傅里叶—梅林矩。经过归一化处理后的雅克比—傅立叶矩具有平移、尺度、旋转、灰度多畸不变性。从归一化图像重建误差、噪声灵敏度等方面对雅可比—傅里叶矩的图像描述能力进行了研究,结果显示雅可比—傅里叶矩在各种参量选择的情况下,具有良好的图像描述和抗噪声能力,是一个优良的图像特征。以前的研究多为单纯构建某一个函数,而现在将其扩展为构建一个函数族,开阔了矩描述量的研究视野。     

几种边界特征描述方法的比较研究

在图像处理与模式识别领域经常根据物体的边界来识别物体。对物体边界特征描述的三种主要方法,即基于极半径函数的傅里叶描述子、边界序列矩和改进的不变矩进行了比较研究,给出了实验结果和结论。     

Robust image hashing using invariants of Tchebichef moments

The content authenticity is critical for secure transmission of multimedia information. As a promising solution, perceptual image hashing has gain great attention. In this paper, we develop a novel algorithm for generating an image hash based on invariants of radial Tchebichef moments. The idea is justified by the fact that the radial Tchebichef moments represent the image under the orthogonal kernel, which has the desirable qualities of orthogonality and robustness. The hash values are achieved by adaptive quantization of the invariants of radial Tchebichef moments, then the random Gay code is applied in the discrete–binary conversion stage to enhance the expected discriminability. Experiments are conducted to show that the proposed hashing algorithm has superior robustness and discrimination performance compared with other state-of-the-art algorithms, in terms of receiving operating characteristic (ROC) curves.     

远程采集模糊无人机图像的特征识别方法研究

地面对空中无人机的视觉识别中,由于无人机的飞行速度、角度呈现非线性变化。使得采集的疑似图像存在特征模糊、衰退等问题,传统的模式识别方法无法提取无人机图像的主要特征,极大程度上降低了图像的识别概率。提出一种引入球面谐波基图像特征细分的无人机识别算法,建立球面谐波基图像识别模型,利用无人机图像的球面谐波基图像近似率,对模糊图像的差异特征进行依次识别。实验结果表明,利用改进算法建立的模糊无人机图像差异特征识别模型,具有一定的优越性,提高了无人机识别的准确率。     

一种新型不变矩在图像识别中的应用

给出了仿射变换的定义,并在此基础上建立了一种新的仿射变换途径——多尺度自卷积变换。构造了一组新的仿射不变矩(MSA矩),分析了其性质,给出了它的明确表达式及其具体实现方法。分别应用最小距离法、k-近邻法和相似性测度三种识别方法进行了仿真实验。结果表明,MSA矩对平移、比例、旋转、拉伸和扭曲等仿射变换均具有很好的不变性,与典型的Flusser矩相比,稳健性好、识别率高、抗噪声干扰能力强,具有一定的应用价值。     

基于光学相关的多目标检测

运用傅里叶变换相关识别技术实现了无畸变、有平移多目标的检测。在此基础上讨论了图像边缘信息和非边缘信息对相关性能的影响。运用分数傅里叶变换相关识别技术实现了目标图像存在尺度、旋转畸变而无平移变化时的目标检测。结果表明,分数傅里叶变换的级次选择对目标图像的旋转和尺度畸变有一定的补偿作用。在分数傅里叶变换相关中,通过简单的级次调节可实现一定范围内的旋转、尺度畸变不变识别。     

基于形状特征的水声图像小目标识别方法

为了抑制背景噪声,提高目标识别准确率,该文提出一种基于形状特征的水声图像小目标识别方法。对含有目标的水声图像进行非局部均值去噪处理后,使用OTSU算法自适应选取阈值对去噪图像进行二值化分割,结合形态学处理获得分割后的目标区域;提取目标区域的矩形度、圆形度、几何不变矩等各项形状参数,将目标的特征向量输入随机森林分类器实现对目标形状的识别。在仿真和实测数据集上分别进行了实验,结果表明,该方法对水声图像中的目标具有较高的识别率,可以实现不同高斯噪声背景下的目标识别,相较于其他方法在识别率上有一定提高。     

Shift- and scale-invariant pattern recognition using morphological phase-only correlation

We propose a non-linear correlation to realize scale-invariant recognition, called shift- and scale-invariant morphological phase-only correlation (SSIMPC), which is a combination of radial harmonic filter (RHF) and morphological phase-only correlation (MPC). We form SSIMPC using phase-only RHFs (PORHF) instead of common RHFs in every sub-correlations of morphological radial harmonic correlation (MRHC). Computer simulation results demonstrate that the performance of SSIMPC is better than that of MRHC in discrimination and robustness against non-Gaussian noise, such as salt-and-pepper noise and cluttered background, and above of all, SSIMOC realizes approximately full scale-invariance. In addition, we devise an optoelectronic scheme, which is identical to the setup for MPC, to implement SSIMPC.     

Regeneration of Form by Fourier Technique and its Application in Powder Characterization

Conventional shape parameters have serious limitations because of their inability to capture finer surface morphology as well as to regenerate the particle form. This problem has been successfully circumvented by applying Fourier analysis technique to the projected image of the particle profile to obtain Fourier coefficients which have been shown to be size inveriant and insensitive to starting point for digitization. Synthesis technique to regenerate the particle profile from Fourier coefficients has revealed that gradual addition of higher order harmonics brings the regenerated form very close to its original counterpart. The regeneration of the form of particles of different shapes as well as application of few strategy to harmonic amplitudes have amply demonstrated the capability of the technique as a potential tool for characterization of powders.     

A New Method for Rotation and Brightness Invariant Pattern Recognition Based on Modified Multiple Circular Harmonic Expansions

A new method for rotation and brightness invariant pattern recognition was proposed by applying multiple circular harmonic expansions to the joint transform correlator. The amplitudes of the multiple orders of circular harmonic expansions made from a detecting image were synthetically modified to respond to the same auto-correlation peaks. These modified circular harmonic expansions were arranged in the input plane as reference patterns together with an arbitrary target pattern, and the correlation signals between them were calculated in the subtracted joint transform correlator. The fraction of the correlation-peak intensities between the target and the references were extracted as a new discrimination parameter. This new parameter performs pattern recognition under rotation and brightness invariance with good discriminability. Its high discriminability has been proved in computer simulations using the face image patterns of many individuals.     

Optical reconstruction from projections via deconvolution

E.W. Hansen and J.W. Goodman have reported recently in this journal (Vol. 24 (1978) p. 268) on optical reconstruction from projections via circular harmonic expansion. Here, the reconstruction problem is derived directly as a deconvolution. Thus the necessary optical processing filter is simply the inverse of the Fourier transform of the convolving point response in the suitably distorted projection image.     

基于多普勒天气雷达数据的中层径向辐合自动识别及其与强对流天气的相关性研究

本文提出了一种从多普勒天气雷达单仰角径向速度图中的正-负速度区域对入手自动识别一类强对流天气的中层径向辐合的方法.首先,根据雷达探测机理,解释了由气流形成的辐合场一定会在径向速度图中形成局部最大正速度区域和局部最大负速度区域的特点,然后构建正-负速度区域对的识别算法,通过来自多个单仰角径向速度图的区域对配准和相关信息,判断雷达径向速度图中是否存在着中层径向辐合,推算辐合的强度、延伸厚度等重要参数,确定得到最优垂直剖面图的剖切方位.经对客观存在显著中层径向辐合的384例样本和不存在明显中层径向辐合的356例样本的测试,本文方法对显著中层径向辐合的识别率达到100%、误识率为0,与人工方法相比,本文方法把对中层径向辐合的识别时间从分钟级缩短到秒级,同时自动给出其强度、高度、厚度、位置等丰富的定量信息和恰当的可视化垂直剖面图.利用给出的参数信息进行了与强对流天气的相关性研究,验证了中层径向辐合与对流行地面大风的较强相关性,同时发现,最强中层径向辐合的高度对区分强冰雹与强降水、中层径向辐合的强度对估计冰雹尺寸具有良好的指示性.     

Gamma畸变的相位误差模型与Gamma标定技术

Gamma畸变是数字相移测量技术的主要误差源。以通用的均匀步长相移技术为对象,分析了Gamma畸变对相位计算的影响,建立了相位误差与谐波系数的关系模型,证明了各阶谐波系数在Gamma值影响下的递推公式,进而提出了基于离散傅里叶变换的Gamma标定技术。基于此Gamma标定值,通过Gamma预矫正降低相位误差。实验结果表明,标定的Gamma值在整个像平面具有较强的稳定性,Gamma矫正后使相位误差减小了77.5%,在曲面测量的结果中,水波纹明显得到抑制,曲面质量得以提高。     

Fractional Fourier plane image encryption technique using radial hilbert-, and Jigsaw transform

A new method for image encryption using integral order radial Hilbert transform (RHT) filter in the fractional Fourier transform (FRT) domain has been proposed. The technique is implemented using the popular double random phase encoding method in the fractional Fourier domain. The random phase masks (RPMs), integral orders of the RHT, fractional orders of FRT, and indices of the Jigsaw transform (JT) have been used as keys for encryption and decryption. Simulation results have been presented and the schematic representation for optical implementation has been proposed. The mean-square-error and signal-to-noise ratio between the decrypted image and the input image have been calculated for the correct as well as incorrect orders of the RHT. Effect of occlusion and noise on the performance of the proposed scheme has also been studied. The robustness of the technique has been verified against attack using partial windows of the correct random phase masks. Similar investigations have also been carried out for the chosen-, and the known-plain-text attacks.     

Compressed imaging system with linear sensors

This Letter presents a new approach for imaging using a linear (vector) sensor. It exploits the fact that visual information within common human intelligible images may be compressed within only a partial set of radial strips of its Fourier domain. We present two imaging schemes, one coherent and the other incoherent, that capture the partial set of radial strips of the object Fourier domain. Two main advantages of the new approach are that the image is captured directly in a compressed form and that the acquisition time is shorter compared with conventional scanning imaging systems.     

三维显示中基于彩色叠栅条纹的视差障栅参数设计

为确定三维显示中视差障栅(PB)的最优参数,提出一种基于彩色叠栅条纹的PB参数设计方法。液晶显示器(LCD)和PB简化为相应参数的黑白光栅模型,而具有特定参数的特定辐射光栅作为周期和倾角变化的PB模型,通过序数方程和傅里叶分析方法分析特定辐射光栅和LCD单色等效黑白光栅的叠加图样,由此推断出不同PB周期的低频主导条纹频率组项,然后利用傅里叶理论的部分和提取方法(PSE)分析不同PB周期的低频主导叠栅条纹的强度轮廓,得出PB的最佳节距和相应倾斜角度。实验结果表明,所设计的PB三维显示系统没有出现降低像质的叠栅条纹,具有较好的三维显示效果。     

基于载频条纹相位分析的畸变测量和校正

为了对图像畸变进行测量和校正,利用纵向朗奇(Ronchi)载频条纹作为测量模板,通过成像系统获取畸变光栅条纹.运用傅里叶变换对畸变条纹图像进行频谱分析、滤波提取基频信息,直接从畸变图像的中心无畸变区域提取理想条纹像信息.通过相位分析提取包裹相位并解包,获得畸变光栅条纹的径向畸变相位分布规律.将该分布规律转化为径向位置畸变分布规律,并结合双线性插值灰度重建对畸变图像进行校正.实验结果证明该方法是有效的.     

Applying backpropagation neural networks to fringe analysis

Advances in image processing and optics technology, allied to the development of algorithmic techniques such as the fast Fourier transform and phase stepping, have allowed automatic fringe analysis to be successfully applied to many problems in visual inspection and noncontact surface measurement. However, when confronted with complicated or noisy images the algorithmic techniques tend to be less successful, implying an alternative approach may be necessary. Neural networks offer such an alternative. They have already been applied with some success to such conceptually similar pattern recognition problems, as the classification of fingerprints, the recognition of facial expressions and the identification of hand-written characters. Here, neural networks are applied to two simple fringe analysis problems. Firstly, to find the radius of a one-dimensional curved surface from its simulated fringe projection intensity distribution and, secondly, to identify four lensshaped objects of different radii of curvature from real fringe patterns obtained under different illumination conditions. In the first experiment, the backpropagation and radial basis function network paradigms are compared. In the second case, backpropagation is compared with the fuzzy-artmap paradigm. Performance criteria are the number of training data presentations, the accuracy of interpolation in the simulation experiment and the classification precision for the real data.     

Optical demonstration of shift-, rotation- and limited size-invariant pattern recognition using a circular harmonic-synthetic discrimination function filter

Optical means of performing pattern recognition possessing shift-, rotation- and limited size-invariance is demonstrated. The experimental set-up is a standard Fourier transform correlator with a binary circular harmonic (CGH) filter. The filtering function of the filter is obtained by the synthetic discriminant functions (SDF) method. In the process of synthesizing the filtering function, circular harmonic functions are chosen as the basic functions. Both optical results and computer simulation results demonstrate that the filter is able to recognize the target in every rotational orientation and in the relative size range of 1–1.82.