结合几何代数改进的SIFT桥梁裂缝图像拼接算法

针对桥梁裂缝图像精度要求高,拼接质量受原图像亮度变化大、噪声干扰严重和对比度低的影响,提出了一种结合几何代数改进的SIFT桥梁裂缝图像的新型拼接算法。对SIFT算法进行了两方面的改进:一是通过几何代数空间的表示形式提取了待拼接图像的色度图像,克服了SIFT算法中色度信息丢失的不足;二是改进了SIFT算法对灰度图像建立尺度空间的方法,构建了新的可适用于多光谱图像的高斯滤波和卷积运算,确定了尺度空间。通过几何代数DoG空间检测特征点并进行预匹配。使用改进的RANSAC算法对匹配结果进行修正,完成了图像之间的精确拼接。实验结果表明,所提算法的性能优于SIFT算法,提取的特征点对数量提高了近10%;拼接过程中未产生位错现象,最终拼接结果满足桥梁裂缝图像的精度要求。     

基于特征点自动匹配的图像拼接方法

提出了一种基于特征点自动匹配的图像拼接算法,采用改进的Harris算子提取特征点,保证了提取的效率和精度,根据互相关的双向匹配实现对应特征点的自动匹配,从而建立参考图像与当前图像的对应点对,最后采用最小二乘方法得到图像间的全局变换参数,实现图像的拼接。     

基于SIFT特征的彩色图像拼接方法研究

针对传统特征提取拼接算法在复杂图像中配准过程中出现的过多误匹配,导致拼接后图像出现鬼影、模糊等问题,从而影响拼接图像的质量,提出一种改进的SIFT配准算法。在对目标图像提取SIFT特征后,利用SIFT描述子的尺度以及梯度方向信息建立最小邻域匹配剔除误匹配点,之后利用局部均方根误差(RMSE)评价映射矩阵与RANSAC算法相结合,迭代出精确变换模型。在对图像进行几何矫正后,提出一种自适应的混合线性算法对重合区域图像变换至HIS颜色空间进行图像拼接,最后得到平滑无缝的完整彩色全景拼接图像。实验结果证明,该算法在拼接复杂场景并且重合区域不多时仍有较好的准确性及稳定性。     

基于光场深度序列的大视野图像拼接算法

针对现有光场深度重建算法只能获取单一视角深度信息,三维建模应用受限的问题,提出一种大视野图像拼接算法。该算法基于三维光场重建的深度图像序列,采用双边滤波及插值算法对图像进行去除噪声,利用基于边缘曲率极值的角点检测算法,并通过最小曲率偏差的列间度量匹配实现了图像之间的精确配准,采用改进的加权平均方法对重叠区域进行融合操作,既保留图像细节又拓展图像视野。实验结果表明,该算法能够有效实现光场深度图像序列的平滑拼接,有利于形成大视野深度数据,为大场景的三维重建奠定了基础。     

基于快速傅里叶变换和积分图的快速相关匹配

为克服传统相关匹配算法计算复杂度过大的缺点,基于快速傅里叶变换和积分图提出了一种快速相关匹配算法.在不改变传统相关匹配计算结果的前提下,该算法通过对相关系数公式的分析和化简,用快速傅里叶变换计算模板图像和基准图像的相关,并采用积分图计算基准图像灰度值及其能量的窗口积分,使相关匹配计算量与模板图像大小近似无关,从而大大降低了计算复杂度.不同光照条件下图像匹配实验和复杂环境下目标跟踪实验结果表明,该算法具有很强的抗干扰能力,其执行效率比传统相关匹配提高1～2个数量级.     

基于系统论匹配准则的尺度不变特征变换的图像自动拼接研究

基于特征点的图像拼接不仅不易受光照、旋转等因素影响,而且还有利于提高速度,是目前图像拼接的主流方向。尺度不变特征变换(SIFT)是目前比较成熟的一种角点检测算法,但是其特征点匹配问题仍然没有得到很好的解决。从系统相似的角度出发,提出了基于系统相似论的匹配准则,并将新匹配准则与传统的匹配准则进行了对比,指出了传统匹配准则存在的问题,测试了新准则匹配的精度和速度,分析了新准则能够取得更高精度的原因,并据新准则成功地检测了待拼接图像中的匹配角点,然后用改进的样本一致性算法计算出仿射变换的投影矩阵,并用Levenberg-Marquardt(L-M)算法对其求精,最终实现了图像的自动拼接。最后给出了全景相机图像的拼接结果,并对新的匹配准则给出了进一步的分析讨论。     

基于灰度相关的高分辨率红外图像重建

介绍了提高空间分辨率的受控和不受控的微扫描法。基于空间灰度相关分布特性,提出了一种微扫描误差修正算法,该算法利用灰度相关特性匹配低分辨率图像的位置关系。为准确评估成像质量,利用高精度METS-L型红外测试系统对MTF进行了测试。测试结果显示,该算法能较好抑制微扫描位移误差引起的图像模糊,有效提高了空间分辨率。     

基于静态小波分解和能量函数优化的图像拼接

提出了一种基于静态小波分解和能量函数优化的全景图拼接算法.该算法首先使用静态小波分解初始图像,并利用序贯检验法对高频图像快速搜索计算粗略匹配序列,在此基础上对初始图像细致搜索计算精确匹配点;对于传统图像融合处理中采用的线性加权函数通常引起的重叠区域模糊问题,采用包含图像梯度的能量函数计算高频图像的全局最优融合因子进行融合,而低频图像则采用线性因子融合.实验表明,该算法在降低匹配运算时间的同时又能获得满意的视觉效果.     

一种改进的块匹配运动目标识别算法

鉴于传统的块匹配法在图像特征提取、匹配以及搜索等方面有诸多不足,复杂的边缘提取处理与基于灰度的匹配准则都使得系统的处理速度与精度降低,为了使识别处理更加快速精确,提出了一种改进的块匹配识别算法。通过拉普拉斯算子对图像特征进行增强,提高图像的特征信息;实验表明,针对320像素240像素大小的图像,改进算法的特征增强计算时间在Matlab中平均约为0005 4 s。     

一种基于单元信息熵的目标匹配改进算法

为解决图像信息熵无法有效进行图像匹配的问题,将图像单元信息熵和投影特征相结合,定义了图像单元信息熵,并提出了一种基于单元投影信息熵的图像匹配方法.在单元信息熵的基础上,在各个单元格内进行单元信息熵投影计算,然后按照一定的测度进行计算,从而实现图像的匹配.采用网格分层的搜索算法,加快搜索速度,提高其工程实用性.实验证明:该算法具有良好的抗几何失真能力和抗辐射失真的能力,以及很好的抗噪声干扰的能力,可以准确的进行目标匹配.     

基于蚁群算法的二维最大熵分割算法

由于二维最大熵分割法不仅考虑了像素的灰度信息,而且还充分利用了像素的空间邻域信息,因此能够取得较好的分割效果.但是,该方法的计算量巨大,不利于红外图像的快速处理.蚁群算法于20世纪90年代初提出,是受到蚁群集体行为的启发而提出的一种基于种群的模拟进化算法,属于随机搜索算法.该算法已经成功应用于旅行商等离散问题.将蚁群算法应用于二维最大熵法,提出了基于蚁群算法的二维最大熵分割算法.与传统的穷尽搜索法相比,求解速度提高了60倍左右.仿真实验表明,该方法快速、简单、有效.     

基于改进的二维指数熵及混沌粒子群的阈值分割

鉴于现常用的灰度级—平均灰度级二维直方图区域划分存在明显的不足,提出了基于灰度级—梯度二维直方图的指数熵阈值选取方法,给出了基于改进的二维直方图的指数熵阈值选取公式,并利用混沌粒子群优化算法寻找最佳分割阈值,采用递推方式降低迭代过程中适应度函数的计算代价。实验结果表明,与现有的有关算法相比,该方法不仅使分割后的图像区域内部更均匀、边界形状更准确、特征细节更清晰,而且使计算效率及粒子群的收敛精度得到提高。     

一种基于粒子群优化的高斯混合灰度图像增强算法

提出一种采用粒子群优化(PSO)的高斯混合灰度图像增强算法。该算法首先采用高斯混合模型(GMM)对输入图像的灰度直方图建模,并采用模型中高斯成分的有效交点来分割直方图。随后,该算法将每个直方图区间的灰度值转换到合适的输出区间,生成增强后的灰度图像,其中转换函数由输入直方图区间的高斯成分和累积分布经过粒子群优化后的参数决定。实验结果显示,该方法生成的图像视觉效果较好,对原图像和纹理细节丰富图像分别进行图像增强,增强后的图像信息熵分别是4.746 6和7.952 6,灰度平均梯度为6.970 6和37.386 1。     

Image enhancement via Median-Mean Based Sub-Image-Clipped Histogram Equalization

This paper presents a robust contrast enhancement algorithm based on histogram equalization methods named Median-Mean Based Sub-Image-Clipped Histogram Equalization (MMSICHE). The proposed algorithm undergoes three steps: (i) The Median and Mean brightness values of the image are calculated. (ii) The histogram is clipped using a plateau limit set as the median of the occupied intensity. (iii) The clipped histogram is first bisected based on median intensity then further divided into four sub images based on individual mean intensity, subsequently performing histogram equalization for each sub image. This method achieves multi objective of preserving brightness as well as image information content (entropy) along with control over enhancement rate, which in turn suits for consumer electronics applications. This method avoids excessive enhancement and produces images with natural enhancement. The simulation results show that MMSICHE method outperforms other HE methods in terms of various image quality measures, i.e. average luminance, average information content (entropy), absolute mean brightness error (AMBE) and background gray level.     

MPPC low-level-light imaging enhancement algorithm based on sub-window box filtering北大核心CSCD

针对多像素光子计数器(MPPC)进行微光成像时,图像受光照不足和噪声影响出现的图像亮度低、对比度差、边缘模糊等问题,提出一种基于子窗口盒式滤波的自适应微光图像处理算法。为了减少算法运行时间的同时突出图像的边缘细节信息,利用子窗口盒式滤波器对图像进行分层得到基础层和细节层;对基础层图像采用自适应阈值直方图均衡化拉伸对比度,细节层图像采用自适应增益控制方式进行增强;根据基础层图像中有效灰度值个数占总灰度的比值自适应确定融合系数,将基础层图像与细节层图像融合得到增强后图像。通过微光实验平台设置3组不同照度的微光环境进行实验仿真,验证了本文算法在保持边缘信息和增强细节方面获得了更好的效果。实验结果表明本文算法在标准差、信息熵、平均梯度等客观评价方面优于改进前算法,提升了微光图像的成像效果。     

二维广义模糊熵图像阈值分割法

针对一维广义模糊熵不能有效处理含噪图像的分割问题,在二维灰度直方图上定义了图像的二维隶属度函数,提出了二维广义模糊熵阈值分割法.该方法不仅考虑了图像的点灰度值,同时考虑了图像像素的邻域平均灰度值,能更好地利用图像中的信息.为了提高二维广义模糊熵阈值法的运行速度、解决参量选取问题,结合粒子群优化搜索方法,设计了嵌套式的优化过程.实验表明,二维广义模糊熵阈值分割法对噪音图像有更好的适应性.     

Research on Color Image Encryption Algorithm Based on Bit-Plane and Chen Chaotic System

In response to the problems of high complexity and the large amount of operations of existing color image encryption algorithms, a low-complexity, low-operation color image encryption algorithm based on a combination of bit-plane and chaotic systems is proposed that is interrelated with plaintext information. Firstly, three channels of an RGB image are extracted, and the gray value of each pixel channel can be expressed by an eight-bit binary number. The higher- and lower-four bits of the binary gray value of each pixel are exchanged, and the position of each four-bit binary number is scrambled by a logistic chaotic sequence, and all the four-bit binary numbers are converted into hexadecimal numbers to reduce the computational complexity. Next, the position of the transformed image is scrambled by a logistic chaotic sequence. Then, the Chen chaos sequence is used to permute the gray pixel values of the permuted image. Finally, the gray value of the encrypted image is converted into a decimal number to form a single-channel encrypted image, and the three-channel encrypted image is synthesized into an encrypted color image. Through MATLAB simulation experiments, a security analysis of encryption effects in terms of a histogram, correlation, a differential attack, and information entropy is performed. The results show that the algorithm has a better encryption effect and is resistant to differential attacks.     

基于灰度变换的红外图像实时分割算法

针对高灰度等级红外数字图像提出了一种基于灰度变换的实时分割算法.通过实时分析原始图像中几个小区域的灰度直方图,自动确定分割阈值,将原始图像通过灰度变换实时转换为低灰度等级图像.该算法能够将高灰度等级的红外数字图像实时显示在普通显示器上,并且能够有效抑制背景并保留目标灰度信息.由于变换后的图像灰度等级较低而且背景及噪音信息较弱,极大地提高了红外跟踪器的性能.     

Macroscopic Time Evolution and MaxEnt Inference for Closed Systems with Hamiltonian Dynamics

MaxEnt inference algorithm and information theory are relevant for the time evolution of macroscopic systems considered as problem of incomplete information. Two different MaxEnt approaches are introduced in this work, both applied to prediction of time evolution for closed Hamiltonian systems. The first one is based on Liouville equation for the conditional probability distribution, introduced as a strict microscopic constraint on time evolution in phase space. The conditional probability distribution is defined for the set of microstates associated with the set of phase space paths determined by solutions of Hamilton’s equations. The MaxEnt inference algorithm with Shannon’s concept of the conditional information entropy is then applied to prediction, consistently with this strict microscopic constraint on time evolution in phase space. The second approach is based on the same concepts, with a difference that Liouville equation for the conditional probability distribution is introduced as a macroscopic constraint given by a phase space average. We consider the incomplete nature of our information about microscopic dynamics in a rational way that is consistent with Jaynes’ formulation of predictive statistical mechanics, and the concept of macroscopic reproducibility for time dependent processes. Maximization of the conditional information entropy subject to this macroscopic constraint leads to a loss of correlation between the initial phase space paths and final microstates. Information entropy is the theoretic upper bound on the conditional information entropy, with the upper bound attained only in case of the complete loss of correlation. In this alternative approach to prediction of macroscopic time evolution, maximization of the conditional information entropy is equivalent to the loss of statistical correlation, and leads to corresponding loss of information. In accordance with the original idea of Jaynes, irreversibility appears as a consequence of gradual loss of information about possible microstates of the system.     

Canny edge detection enhancement by general auto-regression model and bi-dimensional maximum conditional entropy

This paper proposes a novel Canny algorithm without manually setting parameters. Adaptive filter design, implementation and automatic calculation of low and high thresholds are studied in this paper. A general auto-regressive model is deduced that uses only uniform expression for both the linear and non-linear autoregressive model based on Weierstrass theory. Moreover, the bi-dimensional expression of the model is deduced by using bi-vectors instead of scalar parameters. The Generalized M-estimator is chosen for the new model. An adaptive filter is implemented based on the general auto-regression model and simulations are carried out. Gray entropy mathematical model is established according to the gray level-gradient co-occurrence matrix of image and the simulated annealing algorithm is used to solve the gray entropy model. Experiments are done on the worldwide datasets to evaluate the performance of our method. Results demonstrate the superiority of our method compared with the best parameter values method and standard Canny, especially when images are polluted by mixed noises containing Gaussian noise, Poisson noise and impulse noise.