结合邻域信息粒子群聚类用于SAR图像变化检测

SAR图像变化检测可以转化为对差异图的聚类问题。由于 SAR 图像本身容易受到斑点噪声干扰,为提高聚类效果提出了一种结合邻域信息的自适应粒子群聚类算法。该方法在模糊 C 均值原目标函数基础上,引入中心像素的邻域信息,并通过自适应粒子群的全局搜索来优化聚类中心。该方法还引入了自学习算子即粒子编码中的中心像素的隶属度,能够向其相邻像素的隶属度学习,并据此修正自身的隶属度值相关。实验结果表明,与模糊C均值和量子免疫克隆聚类算法相比,该方法利用了像素的邻域信息,从而增强了抗噪性能。与模糊局部信息C均值算法相比,该方法对图像细节保持能力较强,运行时间也较少。     

基于局部相似性测度的SAR图像多层分割

针对谱聚类算法计算复杂度高,不适用于合成孔径雷达图像分割的问题,利用谱聚类算法与权核k均值之间的等价性,提出一种基于局部相似性测度的SAR图像多层分割算法．首先提取图像中每个像素的小波纹理特征,利用每个像素点的纹理特征计算各自的局部尺度参数,进而构造像素点之间的邻接关系,然后利用最近邻规则对此邻接关系进行逐层合并,进行基础聚类和逐层细化实现像素点聚类,最终得到图像的分割结果．对人工纹理图像和SAR图像的分割结果表明了新算法避免了传统谱聚类算法对尺度参数的敏感性,获得了更优的分割性能．     

基于深度学习的遥感影像变化检测方法

为了获得可靠的训练样本及提高遥感影像变化检测的精度,提出基于深度学习的遥感影像变化检测方法. 采用结构相似性方法（SSIM）选取纹理特征（灰度共生矩阵法）,通过融合变化向量分析（CVA）方法获取不同时相遥感影像差异图（DI）及纹理特征差异图获得差异影像,并采用构造的变分去噪模型对差异影像进行去噪. 利用频域显著性方法获取去噪差异影像的显著性图,通过模糊c-均值（FCM）算法对粗变化检测图（对显著性图选取阈值获得的）进行预分类（变化类、未变化类及未确定类）. 将从遥感影像上提取的变化像素和未变化像素的邻域特征引入深度神经网络模型进行训练,并利用训练好的深度神经网络模型对差异影像进行变化检测,得到最终的变化检测图. 对3组遥感影像数据集进行变化检测实验,结果表明本研究方法的变化检测精度高于其他比较方法.     

模糊C-均值聚类图像分割算法的一种改进

针对传统模糊C-均值聚类算法对含噪图像分割时未充分考虑空间信息的问题,提出一种改进的模糊C-均值聚类算法,将图像的局部和非局部两种空间信息引入到模糊C-均值聚类算法的目标函数中,以使两种空间信息在含噪图像分割中发挥互补作用。将改进算法应用于不同含噪图像的分割实验,结果表明图像像素的均方误差均比改进前有所降低。     

高效的彩色图像塔形模糊聚类分割方法

这里提出了一种高效的基于模糊c均值(FCM)聚类的彩色图像分割方法,它利用塔形数据结构对彩色图像进行多层分割。通过对一个彩色图像的分割处理,结果表明,文中所用方法的计算时间仅是用FCM聚类算法而不用塔形进行分割下所需计算时间的十三分之一。     

基于MAP和RF的无监督SAR图像变化检测

基于形态学属性断面(MAP)和随机森林(RF)分类器,提出了无监督合成孔径雷达(SAR)图像变化检测方法.首先,利用MAP算法提取差异图像的几何结构特征,构造深入描述图像结构化信息的特征向量空间;然后,在结合阈值法和偏移因子自动选取训练样本的基础上,用RF分类器在多维特征空间中对图像进行变化与否的判别;最后,利用数学形态学方法对虚警进行滤除.实验结果表明,与传统的基于阈值的变化检测方法相比,该方法不仅能很好地检测出变化区域,而且具有更高的检测精度.     

加权空间函数优化FCM的SAR图像分割

传统模糊c-均值聚类算法没有考虑图像像素空间信息特征,在应用于合成孔径雷达图像分割时,由于合成孔径雷达图像中斑点噪声的影响,通常不能得到正确的分割结果．基于此问题提出加权空间隶属度和加权空间函数并应用于c-均值聚类算法,加权空间隶属度是多尺度条件下空间各相邻像素的位置和强度信息的加权隶属度值,加权空间函数中各加权空间隶属度的影响系数由自适应遗传算法优化,最终的隶属度值由加权空间函数修正．由于在这种聚类过程中融入了优化的空间信息,因此弱化了斑点噪声的影响,提高了分割精度．这种算法应用于实际合成孔径雷达图像分割实验,结果表明此算法对初始分类结果不敏感,具有较强的抗噪性能,改善了SAR图像的分割结果．     

视觉特征方向流邻域加权PCM的SAR图像分割

应用聚类方法对合成孔径雷达(SAR)图像进行分割时,受SAR图像斑点噪声的影响,在聚类过程中应当既要考虑聚类原型的空间自适应性,又要考虑像素间强相关性,这就要求聚类隶属度与空间信息的有效结合．针对此问题提出了基于方向流场构建自适应邻域空间加权可能性c-均值聚类(PCM)算法,通过可操纵小波变换和视觉特征预测编码模型建立方向流场,用基于方向流场的Markov随机场(MRF)描述当前像素与其邻域像素间的相互关系,在像素的聚类隶属度估计中直接引入这种邻域信息,使聚类隶属度得到有效修正．实验表明这种方法在抑制噪声的同时可以保留图像的细节信息,对SAR图像有较好的分割结果．     

Optimal multilevel thresholding based on molecular kinetic theory optimization algorithm and line intercept histogram

Among all segmentation techniques, Otsu thresholding method is widely used. Line intercept histogram based Otsu thresholding method(LIH Otsu method) can be more resistant to Gaussian noise, highly efficient in computing time, and can be easily extended to multilevel thresholding. But when images contain salt-and-pepper noise, LIH Otsu method performs poorly. An improved LIH Otsu method(ILIH Otsu method) is presented, which can be more resistant to Gaussian noise and salt-and-pepper noise. Moreover, it can be easily extended to multilevel thresholding. In order to improve the efficiency, the optimization algorithm based on the kinetic-molecular theory(KMTOA) is used to determine the optimal thresholds. The experimental results show that ILIH Otsu method has stronger anti-noise ability than two-dimensional Otsu thresholding method(2-D Otsu method), LIH Otsu method, K-means clustering algorithm and fuzzy clustering algorithm.     

模糊逻辑和谱聚类的字符图像分割

为了从复杂背景中有效分离出字符图像,提出了一种基于模糊逻辑的谱聚类字符图像分割方法.利用最大信息熵准则获得模糊函数的参数,将原始图像模糊化;在模糊后的图像上建立像素间的相似矩阵,文本图像的纹理、灰度及像素间的距离是定义相似函数的依据,计算相似矩阵最小特征值对应的特征向量,并对其聚类划分;利用分类后的特征向量对相似矩阵进行划分,进而实现原图像的分割.实验结果表明:本文方法优于一般的阈值化分割方法,能够有效处理背景复杂的自然场景文本图像.     

Road network extraction in classified SAR images using genetic algorithm

Due to the complicated background of objectives and speckle noise, it is almost impossible to extract roads directly from original synthetic aperture radar(SAR) images. A method is proposed for extraction of road net-work from high-resolution SAR image. Firstly, fuzzy C means is used to classify the filtered SAR image unsupervis-edly, and the road pixels are isolated from the image to simplify the extraction of road network. Secondly, according to the features of roads and the membership of pixels to roads, a road model is constructed, which can reduce the extraction of road network to searching globally optimization continuous curves which pass some seed points. Final-ly, regarding the curves as individuals and coding a chromosome using integer code of variance relative to coordi-nates, the genetic operations are used to search global optimization roads. The experimental results show that the al-gorithm can effectively extract road network from high-resolution SAR images.     

Sonar Image Detection Algorithm Based on Two-Phase Manifold Partner Clustering

According to the characteristics of sonar image data with manifold feature, the sonar image detection method based on two-phase manifold partner clustering algorithm is proposed. Firstly, K-means block clustering based on euclidean distance is proposed to reduce the data set. Mean value, standard deviation, and gray minimum value are considered as three features based on the relatinship between clustering model and data structure. Then K-means clustering algorithm based on manifold distance is utilized clustering again on the reduced data set to improve the detection efficiency. In K-means clustering algorithm based on manifold distance, line segment length on the manifold is analyzed, and a new power function line segment length is proposed to decrease the computational complexity. In order to quickly calculate the manifold distance, new all-source shortest path as the pretreatment of efficient algorithm is proposed. Based on this, the spatial feature of the image block is added in the three features to get the final precise partner clustering algorithm. The comparison with the other typical clustering algorithms demonstrates that the proposed algorithm gets good detection result. And it has better adaptability by experiments of the different real sonar images.     

A fast and effective fuzzy clustering algorithm for color image segmentation

A fast and effective fuzzy clustering algorithm is proposed. The algorithm splits an image into n×n blocks, and uses block variance to judge whether the block region is homogeneous. Mean and center pixel of each homogeneous block are extracted for feature. Each inhomogeneous block is split into separate pixels and the mean of neighboring pixels within a window around each pixel and pixel value are extracted for feature. Then cluster of homogeneous blocks and cluster of separate pixels from inhomogeneous blocks are carried out respectively according to different membership functions. In fuzzy clustering stage, the center pixel and center number of the initial clustering are calculated based on histogram by using mean feature. Then different membership functions according to comparative result of block variance are computed. Finally, modified fuzzy c-means with spatial information to complete image segmentation are used. Experimental results show that the proposed method can achieve better segmental results and has shorter executive time than many well-known methods.     

Color image segmentation using mean shift and improved ant clustering

To improve the segmentation quality and efficiency of color image, a novel approach which combines the advantages of the mean shift (MS) segmentation and improved ant clustering method is proposed. The regions which can preserve the discontinuity characteristics of an image are segmented by MS algorithm, and then they are represented by a graph in which every region is represented by a node. In order to solve the graph partition problem, an improved ant clustering algorithm, called similarity carrying ant model (SCAM-ant), is proposed, in which a new similarity calculation method is given. Using SCAM-ant, the maximum number of items that each ant can carry will increase, the clustering time will be effectively reduced, and globally optimized clustering can also be realized. Because the graph is not based on the pixels of original image but on the segmentation result of MS algorithm, the computational complexity is greatly reduced. Experiments show that the proposed method can realize color image segmentation efficiently, and compared with the conventional methods based on the image pixels, it improves the image segmentation quality and the anti-interference ability.     

Mercer Kernel Based Fuzzy Clustering Self-Adaptive Algorithm

A novel mercer kernel based fuzzy clustering self-adaptive algorithm is presented. The mercer kernel method is introduced to the fuzzy c-means clustering. It may map implicitly the input data into the high-dimensional feature space through the nonlinear transformation. Among other fuzzy c-means and its variants, the number of clusters is first determined. A self-adaptive algorithm is proposed. The number of clusters, which is not given in advance, can be gotten automatically by a validity measure function. Finally, experiments are given to show better performance with the method of kernel based fuzzy c-means self-adaptive algorithm.     

彩色图像分割在视频跟踪系统中的应用研究

在视频跟踪系统中,图像分割是进行跟踪的前期处理。只有在视频序列图像中有效的分割出运动目标,才能实现后续对目标的准确跟踪。提出一种混合算法,将一种基于灰度直方图的阈值化分割算法应用到了HSI颜色空间上,利用H进行阈值化,在阈值化之前,先根据R、G、B值对像素进行了筛选;文章应用董立菊等人提出的-种基于RGB空间的K均值聚类算法、一种基于灰度直方图的阈值化算法和混合算法,以目标颜色为特征,对彩色图像进行了分割;针对特定的视频跟踪系统,对各结果进行了比较,得出了结论,找出了较优算法——混合算法效果较理想,能够较有效的分割目标,为后续跟踪工作做好了前期处理工作。     

模糊聚类的遗传算法在图像压缩中的应用

图像压缩的关键在于对图像中相似取样点的选择和处理.用基于图像像素的遗传聚类算法对3D图像分析选择样本,通过FCM算法得到一个有序的像素序列,然后进行聚类获得压缩.实验表明,运用模糊聚类分析法能够有效降低压缩算法的复杂度,并能达到预期的图像压缩效果.     

一种基于最大全变差准则的SAR图像自聚焦算法

研究了一种基于信号全变差的合成孔径雷达图像自聚焦算法．该方法从复图像域出发,通过在距离压缩相位历史域引入相位误差模型,改变图像的聚焦程度直至一维像的全变差达到最大,从而完成误差校正．同最小熵法相比,该自聚焦算法计算量较少,更易实现．     

改进的模糊C-均值算法在图像分割中的应用

针对传统的模糊C-均值算法在图像分割中存在的缺陷,提出了一种基于点密度函数加权的模糊C-均值聚类算法。将图像像素的点密度函数作为权值,并依据类间相关度定义了一个聚类有效性函数用以确定最佳聚类数,结合聚类有效性完成对图像的分割。理论分析和对比试验表明,该算法在一定程度上克服了模糊均值算法的缺陷,在图像分割中具有良好的分类精度。     

Recognition of Spontaneous Combustion in Coal Mines Based on Genetic Clustering

1 Introduction Coal mine fires can be divided into two kinds. One is an exogenous fire; the other is an internally caused fire [1]. Exterior hot objects are the cause of exterior-caused fires, such as fire damp explosions, mechanical friction, an electrical short circuit, joint sparks, etc. Internally caused fires happen in places where is oxygen deficient, such as laneways, goafs and coalholes. Spontaneous combustion of coal may be caused by three necessary conditions: the self- ignition pot…