基于混合幅度-相位检索技术与二维耦合混沌映射的光学图像加密算法

目的为了解决当前光学图像加密算法主要将单色光束直接作用于明文,使其在解密过程中易出现丢失颜色信息等问题。方法文中设计基于混合幅度-相位检索技术与二维耦合混沌映射的光学图像加密算法。首先,提取彩色图像的R, G, B分量;随后,引入Logistic映射与Sine映射,通过对二者进行非线性耦合,形成二维复合混沌映射;利用彩色图像的像素信息来迭代复合映射,获取3个混沌序列,通过构建位置引擎混淆机制,对R,G,B分量进行置乱;基于Logistic映射,利用明文像素生成的初值条件对其进行迭代,输出一个混沌随机掩码;最后,基于幅度-相位截断方法和Gyrator变换,设计混合幅度-相位检索技术,利用单向二进制相位函数和随机掩码,对置乱后的R, G, B分量进行加密,获取相应的检测振幅,再将其进行组合,形成实值函数的加密密文。结果实验结果显示,与当前光学图像加密机制相比,所提算法具有更高的安全性与解密质量,具备较强的抗明文攻击能力。结论所提加密技术具有较高的抗攻击能力,能够安全保护图像在网络中传输,在信息防伪等领域具有较好的应用价值。     

用Gabor滤波实现虹膜图像纹理编码

针对虹膜纹理的幅度和相位信息特征,利用二维Gabor滤波器族可以很好地模拟皮层简单细胞的二维感受野轮廓,能够最大限度的提供图像局部方向和频率信息的特点,提出一种基于二维Gabor函数族对虹膜图像进行子块分解滤波的改进编码算法。实验结果表明,该方法能有效的提取虹膜纹理特征,从而达到虹膜识别的目的。     

基于圆柱衍射与离散余弦变换的图像光学加密算法

目的为了实现多幅图像的同步加密,并增强加密系统的抗破译能力,提出一种基于圆柱衍射域的相位截断与离散余弦变换的多图像光学加密算法。方法首先引入压缩感知(CS,Compress Transform)方法,对输入明文实施压缩;基于离散余弦变换DCT(Discrete Cosine Transform)对压缩明文完成分解,获取相应的DCT系数,形成系数矩阵;构建迭代复数,将每个压缩明文对应的系数矩阵融合为一个复矩阵,通过DCT逆变换,形成一幅组合图像。联合Hilbert变换与波带片相位模型,构建调制掩码;引入圆柱衍射域的相位截断机制,联合调制掩码,对组合图像实施光学加密,获取密文与私钥。结果实验数据表明,相对于已有的多图像同步加密方法而言,所提算法具备更高的加密安全性,密文熵值以及相邻像素间的相关系数分别达到了7.998,0.0012,且具有强烈的密钥敏感性。结论所提加密算法可以抵御网络中外来攻击,在图像信息防伪领域具有一定的参考价值。     

基于方向金字塔分解与稳定几何失真校正的鲁棒图像水印算法

目的为了提高水印算法的抗几何攻击能力,并兼顾较高的鲁棒性与不可感知性,设计一种基于方向金字塔分解与稳定几何失真校正的鲁棒图像水印算法。方法首先,引入方向金子塔,对载体图像完成分解,输出对应的低通与高通子带;将低通子带分割为一系列的非重叠块;根据载体的亮度、纹理与边缘掩码,计算水印嵌入强度,最大程度地平衡水印图像的不可感知性与鲁棒性;设计水印嵌入方法,将经过Arnold映射加密后的水印嵌入到非重叠子块中,通过修改载体的方向金子塔分解系数,获取水印图像;将不同的攻击类型作用于水印图像,建立训练样本;再利用方向金子塔分解训练样本,计算高通子带的高斯-厄米矩能量,将其视为特征矢量;再利用特征矢量对模糊支持向量机完成训练,以预测几何失真参数,准确校正受攻击的水印图像;设计水印检测机制,从水印图像中复原水印。结果实验数据表明,与当前图像水印方案相比,所提算法具有更高的抗几何变换能力,以及较好的不可感知性与鲁棒性,其提取的水印失真度最小,对应峰值信噪比保持在40dB以上。结论所提水印算法具有较高的鲁棒性和视觉隐秘性,在版权保护、信息防伪等领域具有一定的参考价值。     

随机振幅-相位编码虚拟光学加密系统

通过对双随机相位编码光学加密系统中空域的随机相位模板进行置换,提出采用随机振幅-相位模板对图像进行加密的新方法。理论分析表明,当随机振幅板为零均值白噪声且与频域相位板统计独立时,被加密图像的自相关函数为二维冲击函数,表明原始图像被加密为平稳的复随机白噪声,因而可以抵御盲反卷积攻击。采用计算机进行模拟,证实了所提方法的有效性。最后讨论了该方法相对于双随机相位编码系统的优越性:对空域密钥的统计学要求更低、在频域密钥泄露情况下系统鲁棒性更强。但是,由于物理上不存在幅度为负值的光波,因此该系统只能使用数字虚拟的方法来实现。     

一种基于相关积分的互谱WVD目标方位估计方法

针对水下高速运动目标的被动跟踪问题,将魏格纳-威尔分布(Wigner-Ville Distribution,WVD)算法与互谱法相结合,构建了一种矢量信号处理框架下的目标方位估计方法——基于相关积分的互谱WVD算法。该算法利用了矢量水听器声压通道与振速通道信号的相关特性,首先通过计算两者的互相关函数来提取目标信号的特征信息,然后通过短时积分与傅里叶变换将互相关函数从时域转换到频域,最后在频域提取信号的特征参量,并进行方位估计。仿真研究表明,当目标处于远距离且低速运动时,所提算法的方位估计性能与互谱声强法基本一致;而当目标处于近距离且高速运动时,算法的方位估计性能大大优于互谱声强法。     

基于经验模态分解和主成分分析的滚动轴承故障诊断研究

针对传统的振动信号特征提取效率低、诊断时间较长等问题,提出了基于经验模态分解与主成分分析的滚动轴承故障诊断方法。首先利用经验模态分解将振动信号分解为有限个本征模函数和一个残差函数,提取主要的本征模函数能量及其局部平均频率特征,最后将复合特征向量作为主成分分析分类器的输入,完成对故障的识别。实验结果表明:复合特征向量能够有效地反映轴承的运行状态;相比于BP神经网络、支持向量机、K-近邻算法,主成分分析分类法不仅能够准确地识别故障,而且训练时间短、使用方便。     

基于Potrace点跟踪的文创产品逆向设计方法研究

目的 为辅助设计人员进行产品创新设计,特别是针对文创产品设计领域提高设计效率和产品快速迭代的需求,利用产品逆向设计方法,对设计实物二维形态进行快速扫描,获取其外部形态轮廓数据,在此基础之上进行改良或者创新设计,对于快速生成产品创意和设计实践都提供了新的方法和思路。方法 针对逆向设计中,轮廓提取细节缺失、弧度线条生硬问题,本文提出了一种基于Potrace点跟踪的图像矢量路径优化算法用于文创产品的数字化分析及逆向设计。此方法以点跟踪作为基础,先利用位图中的顶点将图像轮廓抽象成多边形路径,然后再根据多边形转角大小利用贝塞尔曲线进行优化,最后通过交叉熵损失函数优化最终路径。结果 实验表明,算法在复杂边缘的矢量化线条过渡好于其他算法,并且算法与目前常见矢量化算法相比较降低了时间复杂度,算法运行时间较其他算法能够提升40%以上。结论 针对Potrace方法,提出了一种新的损失函数优化算法,算法在复杂边缘处过渡优于其他算法,同时算法运行时间大大降低,有利于将此算法用于文创产品逆向设计方法研究中。     

基于显著像素复合矩阵的多图像同步实时加密算法

目的 研究基于显著像素复合矩阵的多图像同步无损实时加密算法。方法 提出引入ZigZag机制,扰乱所有明文像素位置。定义显著像素择取机制, 将置乱后的密文像素分割为显著像素与次要像素, 形成多个显著像素矩阵。设计迭代复数模型, 将这些显著像素矩阵形成复合矩阵, 借助奇异值分解与Logistic映射, 得到矩阵密钥最后构造扩散函数, 利用矩阵密钥对显著像素复合矩阵执行扩散, 获取密文。结果 文中加密机制高度安全, 且无失真, 与当前多图像加密机制相比, 文中算法的加密效率更高,可满足实时性传输需求。结论 文中算法可同时对多幅图像进行实时安全加密传输。     

基于卷积神经网络图像分类的轴承故障模式识别

针对传统的基于数据驱动的机械故障模式识别方法中需要人工构造算法提取特征以及人工构造特征提取算法繁琐的问题,结合卷积神经网络(CNN)在图像特征自动提取与图像分类识别中的广泛应用,提出了一种基于CNN图像分类的轴承故障模式识别方法。首先,利用集合经验模态分解(EEMD)方法对轴承振动信号进行自适应分解并用相关系数对得到的本征模函数分量进行筛选。其次,对筛选得到的本征模函数分量进行伪魏格纳-威利时频分析(PWVD)计算得到信号的时频分布图,并对时频图进行预处理。最后,将轴承15种不同工况预处理后的时频图利用CNN进行特征提取与分类识别。将该方法与同类方法进行了对比,分类正确率提高了4.26%。     

Binary image encryption based on interference of two phase-only masks

Optical image encryption based on interference has attracted a lot of attention recently. The technique employs two pure phase masks derived from the complex field of the image in the Fresnel diffraction domain. The image decryption procedure can be carried out by inverse Fresnel transformation of the summation of two pure phase masks. However, the silhouette of the original image, which is recovered by either of the two phase-only masks, impedes the application of this technique. In this paper, a very simple method for binary image encryption based on interference of two phase-only masks is proposed without any silhouette problem. The binary image in combination with a random phase mask is separated into two phase-only masks directly, and the decryption by summation of the two masks can be performed digitally or optically. In this paper, the encryption and decryption processes are analyzed, after which both the optical simulation and the experimental results based on single-beam holography are given to demonstrate the feasibility of the encryption method. As information nowadays is mainly digitized into binary codes, the proposed encryption method may find applications in the information processing field.     

Optical image hiding with silhouette removal based on the optical interference principle

The earlier proposed interference-based encryption method with two phase-only masks (POMs), which actually is a special case of our method, is quite simple and does not need iterative encoding. However, it has been found recently that the encryption method has security problems and cannot be directly applied to image encryption due to the inherent silhouette problem. Several methods based on chaotic encryption algorithms have been proposed to remove the problem by postprocessing of the POMs, which increased the computation time or led to digital inverse computation in decryption. Here we propose a new method for image encryption based on optical interference and analytical algorithm that can be directly used for image encryption. The information of the target image is hidden into three POMs, and the silhouette problem that exists in the method with two POMs can be resolved during the generation procedure of POMs based on the interference principle. Simulation results are presented to verify the validity of the proposed approach.     

Image encryption based on interference that uses fractional Fourier domain asymmetric keys

We propose an image encryption technique based on the interference principle and phase-truncation approach in the fractional Fourier domain. The proposed scheme offers multiple levels of security with asymmetric keys and is free from the silhouette problem. Multiple input images bonded with random phase masks are independently fractional Fourier transformed. Amplitude truncation of obtained spectrum helps generate individual and universal keys while phase truncation generates two phase-only masks analytically. For decryption, these two phase-only masks optically interfere, and this results in the phase-truncated function in the output. After using the correct random phase mask, universal key, individual key, and fractional orders, the original image is retrieved successfully. Computer simulation results with four gray-scale images validate the proposed method. To measure the effectiveness of the proposed method, we calculated the mean square error between the original and the decrypted images. In this scheme, the encryption process and decryption keys formation are complicated and should be realized digitally. For decryption, an optoelectronic scheme has been suggested.     

基于变参混沌的异位异或图像加密算法

目的 解决混沌序列的混沌性能退化,加密算法不能较好地抵御选择明文攻击等问题。方法 提出一种基于变参混沌系统,采用密文反馈的方式,对明文块依次加密的图像加密算法,并在扩散部分摒弃常规的按位异或扩散方式,提出一种按块异位异或算法。结果 对算法进行仿真的结果表明,文中算法对密钥敏感,测试对象的NPCR值分别为99.59%,99.61%,99.61%,UACI值分别为33．44%,33.46%,33.45%,信息熵分别为7.9872,7.9989,7.9977,相较于其他类似算法,文中算法的综合效果更好,密钥空间为(1016)6+1014×9。结论 密文对明文敏感,能抵抗选择明文攻击,该算法不仅有效解决了问题,还具有较高的安全性能。     

分数傅里叶变换域的彩色图像非对称光学压缩加密

为了提高传统双随机相位编码图像光学加密系统的安全性,并减少其所需要处理的数据量,提出了一种基于压缩感知及量子Logistic混沌映射的彩色图像非对称光学加密方法。针对彩色图像加密过程中所需要处理数据量过大问题,首先利用压缩感知理论减少加密系统所需要处理的数据量,其次,将彩色图像三通道转换为单通道加密来减少数据量。针对传统光学加密系统为线性系统问题,采用基于相位截断的非对称光学加密方法进行加密。针对光学加密系统加密密钥为随机相位板不方便传输问题,利用量子混沌产生系统所需要的随机相位板。结果表明,此算法可以获得较为理想的图像加密和解密效果。     

Coaxial holographic encoding based on pure phase modulation

We describe a simple technique for coaxial holographic image recording and reconstruction, employing a spatial light modulator (SLM) modified in pure phase mode. In the image encoding system, both the reference beam in the outside part and the signal beam in the inside part are displayed by an SLM based on the twisted nematic LCD. For a binary image, the part with amplitude of "1" is modulated with random phase, while the part with amplitude of "0" is modulated with constant phase. After blocking the dc component of the spatial frequencies, a Fourier transform (FT) hologram is recorded with a uniform intensity distribution. The amplitude image is reconstructed by illuminating the reference beam onto the hologram, which is much simpler than existing phase modulated FT holography techniques. The technique of coaxial holographic image encoding and recovering with pure phase modulation is demonstrated theoretically and experimentally in this paper. As the holograms are recorded without the high-intensity dc component, the storage density with volume medium may be increased with the increase of dynamic range. Such a simple modulation method will have potential applications in areas such as holographic encryption and high-density disk storage systems.     

基于QR码和矩阵映射的强鲁棒性信息加密技术

目的为了避免传统加密算法直接将图像信息作为明文加密及加密图像含有明文图像信息而带来被破解的风险,提出一种将伪明文图像加密嵌入宿主图像传输的强鲁棒性信息加密技术。方法首先对明文图像彩色QR码和伪明文图像进行初步加密,然后通过加性原则构造出两者之间的映射矩阵,最后基于小波变换和奇异值分解将加密后的伪图像嵌入到载体图像中。结果仿真结果表明,该算法具有较高的安全性及强大的抗攻击能力。结论算法通过将QR码作为明文图像建立与伪图像之间的关系,并将伪图像加密嵌入宿主图像传输,不仅有效地避免了明文被破解的风险,还具有较强的抗噪声、压缩、裁剪和旋转等恶意攻击能力,表明该加密算法具有良好的安全性和稳健性。     

Image encryption by encoding with a nonuniform optical beam in gyrator transform domains

Based on an optical gyrator transform system, an image encryption algorithm is designed and studied. An original secret image is regarded as the output intensity of the second gyrator transform. A coherent nonuniform optical beam is converted into the input of the first gyrator transform. A Gerchberg-Saxton phase retrieval algorithm is employed for obtaining the compensation phases in the first gyrator transform pair. The compensation phases are regarded as the encrypted image and key in this algorithm. The parameters of the laser beam and gyrator transform can serve as the additional key of encryption method. The decryption process of this encryption algorithm can be achieved with an optical system. Numerical simulations are performed to test the validity and capability of the encryption algorithm.