一种计算二维图象矩的神经网络方法

提出一种计算二维图象矩的神经网络方法,首先,导出一维信号矩与一维Ｈａｄａｍａｒｄ变换之间的关系。通过２Ｎ－一维Ｈａｄａｍａｒｄ变换和极少量加法,移位及乘法运行可计算二维图象矩。然后给出用Ｔａｎｋ－Ｈｏｐｆｉｅｌｄ神经网络实现Ｈａｄａｍａｒｄ变换的方法,并对该网络作了改进。     

基于小波变换和核独立分量分析的遥感图像变化检测

为了进一步提高基于独立分量分析(IndependentComponentAnalysis,ICA)的遥感图像变化检测精确度,并解决ICA分离的图像分量排序不确定问题,提出了基于小波变换和核独立分量分析(KernelIndependentComponentAnalysis,KICA)的遥感图像变化检测方法。首先通过小波变换对遥感图像进行分解,得到由图像的高频分量和低频分量组成的分块向量,然后利用核函数将分块向量映射到高维特征空间中,再在该空间中用ICA方法分离出互相独立的向量,最后根据分离出的向量中高频分量的差异自动分辨出变化分量。文章给出了遥感图像变化检测方法及近年提出的基于主分量分析(PrincipalComponentAnalysis,PCA)、基于ICA、基于KICA三种变化检测方法的试验结果,并进行了分析和定量比较。试验结果表明,文中方法能更好地分离出遥感图像的变化信息,具有更高的精确度,并实现了变化检测的智能化。     

基于Contourlet域Krawtchouk矩和改进粒子群的遥感图像匹配

为了进一步提高遥感图像匹配的精度和运算效率,提出了一种利用Contourlet变 换、Krawtchouk矩和改进粒子群的遥感图像匹配算法。在分别对参考图像和目标图像进行Co ntourlet分解的基础上,引入Krawtchouk矩来提取图像的局部特征,并利用改进的带极值扰 动的简化粒子群优化算法对低分辨率的遥感图像进行匹配操作,然后逐级上推,
最终实现全分辨率情况下遥感图像的匹配。实验结果表明,该算法与目前常用的遥感图像匹 配算法相比,不仅具有更高的匹配精度和运算效率,还有较强的鲁棒性。
     

抗几何攻击的数字图像版权保护算法

为了在保证图像质量的前提下实施图像的版权保护,提出了一种基于Krawtchouk矩和标识信息的零水印算法.该方法利用从图像中提取的特征信息来构造水印,从而达到版权保护的目的.主要步骤为,首先计算原始图像的Krawtchouk低阶矩不变量来构建水印信息,接着将水印信息和密钥信息在第三方登记注册并获得标识信息,最后通过标识信息确认版权所有者.水印提取过程简单,只需计算所得图像的几个低阶Krawtchouk矩不变量.文中给出了实验结果,并与传统几何矩不变量的算法进行了比较. 结果表明,该方法对旋转、缩放、剪切、组合攻击、平移和去掉行等几何攻击以及滤波、JPEG压缩等攻击具有更高的稳健性.     

Ground-Based Cloud Using Exponential Entropy/Exponential Gray Entropy and UPSO

Objective and accurate classification model or method of cloud image is a prerequisite for accurate weather monitoring and forecast. Thus safety of aircraft taking off and landing and air flight can be guaranteed. Thresholding is a kind of simple and effective method of cloud classification. It can realize automated ground-based cloud detection and cloudage observation. The existing segmentation methods based on fixed threshold and single threshold cannot achieve good segmentation effect. Thus it is difficult to obtain the accurate result of cloud detection and cloudage observation. In view of the above-mentioned problems, multi-thresholding methods of ground-based cloud based on exponential entropy/exponential gray entropy and uniform searching particle swarm optimization （UPSO） are proposed. Exponential entropy and exponential gray entropy make up for the defects of undefined value and zero value in Shannon entropy. In addition, exponential gray entropy reflects the relative uniformity of gray levels within the cloud cluster and background cluster. Cloud regions and background regions of different gray level ranges can be distinguished more precisely using the multi-thresholding strategy. In order to reduce computational complexity of original exhaustive algorithm for multi-threshold selection, the UPSO algorithm is adopted. It can find the optimal thresholds quickly and accurately. As a result, the real-time processing of segmentation of groundbased cloud image can be realized. The experimental results show that, in comparison with the existing groundbased cloud image segmentation methods and multi-thresholding method based on maximum Shannon entropy, the proposed methods can extract the boundary shape, textures and details feature of cloud more clearly. Therefore, the accuracies of cloudage detection and morphology classification for ground-based cloud are both improved.     

Edge Detection of River in SAR Modulus Maxima and Improved Image Based on Contourlet Mathematical Morphology

To cope with the problems that edge detection operators are liable to make the detected edges too blurry for synthetic aperture radar （SAR） images, an edge detection method for detecting river in SAR images is pro- posed based on eontourlet modulus maxima and improved mathematical morphology. The SAR image is firstly transformed to a contourlet domain. According to the directional information and gradient information of direction- al subband of contourlet transform, the modulus maximum and the improved mathematical morphology are used to detect high frequency and low frequency sub-image edges, respectively. Subsequently, the edges of river in SAR image are obtained after fusing the high frequency sub-image and the low frequency sub image. Experimental re sults demonstrate that the proposed edge detection method can obtain more accurate edge location and reduce false edges, compared with the Canny method, the method based on wavelet and Canny, the method based on contourlet modulus maxima, and the method based on improved （ROEWA）. The obtained river edges are complete and clear.     

基3FFT与FHT的新算法

本文首先提出了适合于N=3~m的按频率抽取快速Fourier变换(DIF FFT)与按时间抽取快速Fourier变换(DIT FFT)的新算法,分别导出了算法的一般公式,给出了信号流图,分析了计算复杂性。对于实序列的FFT算法也相应作了探讨,然后按照类似的方式提出了适合于N=3~m的按频率抽取快速Hartley变换(DIF FHT)与按时间抽取快速Hartley变换(DIT FHT)的新算法。最后将本文提出的基3FFT算法、基3FHT算法与其他基3算法进行了比较。结果表明,新算法不仅所需运算量少,优于其他基3算法,而且可进行原位计算,结构简单规则,易用硬件或软件实现。本文算法扩展了FFT算法与FHT算法的可适用范围。     

基于变邻域变步长LMS背景预测检测红外小目标

在分析强起伏背景信号的基础上,利用背景局部信号统计特征和目标运动特性,提出了一种基于变邻域变步长LMS自适应背景预测的红外弱小目标检测方法.首先将两类背景交界处像素的邻点按最大类间平均离差准则分成两类,和中心像素点相近的一类构成预测邻域,而背景内部区域采用固定预测邻域;然后提出了一种改进的变步长LMS自适应算法,在所得预测域上进行背景预测,由实际值和预测值相减得到残差图像;最后采用二维Tsallis-Havrela-Charvat熵阈值选取方法对残差图像进行分割,并根据目标运动的连续性和一致性确认真实小目标.针对实际红外图像序列的实验结果表明:该算法能有效地抑制强起伏杂波,增强目标能量,降低虚警率,对强起伏背景下弱小目标具有很好的检测性能.     

基于NMF、ICA和复Contourlet变换的红外小目标检测

针对存在背景干扰和噪声情况下的红外弱小目标检测问题,提出了一种基于非负矩阵分解(NMF)、独立分量分析(ICA)和复Contourlet变换的检测方法。首先通过非负矩阵分解和独立分量分析分别抑制原始图像的背景,得到不同的小目标残差图像;接着采用复Contourlet变换对残差图像进行去噪;再对上述去噪后的小目标残差图像求和,得到了预处理图像;最后提出基于模糊灰度熵阈值选取方法分割预处理图像,从而实现了复杂背景下的红外弱小目标检测。针对红外小目标图像进行了大量实验,并与基于新型Top-hat变换、基于快速独立分量分析的目标检测方法进行了比较,结果表明所提出的方法抗噪性强,具有更为优越的检测性能。     

基于粒子群聚类优化的快速分形图像压缩算法

针对经典分形压缩算法中编码时间过长的问题提出了一种改进算法。将粒子群优化算法应用于聚类中心的求解中,利用聚类优化方法分别对子块和父块进行聚类,匹配时通过类内搜索取代全局搜索,减低了编码时间。仿真实验结果表明,在不影响信噪比和压缩比的前提下,与经典分形压缩算法相比,本文的算法编码速度可提高大约5倍;同近期文献报道的基于K-均值聚类优化的快速分形图像压缩算法相比也有明显的改善。