基于GA-EM算法的GMM遥感影像变化检测方法*

提出了一种基于GA-EM算法的高斯混合模型(GMM)遥感影像变化检测方法。该方法采用主成分分析(PCA)与传统差值法相结合的方式构造差异影像;然后使用N个成分的GMM对差异影像分布进行建模;再利用进化的迭代方法对模型进行自适应参数估计;最后利用贝叶斯准则实现变化和未变化像元分布的变化检测结果。仿真结果表明,该方法对变化目标的检测有效而可靠,具有较大的实用价值。     

A Hopfield Neural Network for Image Change Detection

This paper outlines an optimization relaxation approach based on the analog Hopfield neural network (HNN) for solving the image change detection problem between two images. A difference image is obtained by subtracting pixel by pixel both images. The network topology is built so that each pixel in the difference image is a node in the network. Each node is characterized by its state, which determines if a pixel has changed. An energy function is derived, so that the network converges to stable states. The analog Hopfield's model allows each node to take on analog state values. Unlike most widely used approaches, where binary labels (changed/unchanged) are assigned to each pixel, the analog property provides the strength of the change. The main contribution of this paper is reflected in the customization of the analog Hopfield neural network to derive an automatic image change detection approach. When a pixel is being processed, some existing image change detection procedures consider only interpixel relations on its neighborhood. The main drawback of such approaches is the labeling of this pixel as changed or unchanged according to the information supplied by its neighbors, where its own information is ignored. The Hopfield model overcomes this drawback and for each pixel allows a tradeoff between the influence of its neighborhood and its own criterion. This is mapped under the energy function to be minimized. The performance of the proposed method is illustrated by comparative analysis against some existing image change detection methods.     

基于改进高斯混合模型的运动目标检测方法

传统高斯混合模型(GMM)对于光照突变十分敏感,且收敛速度较慢。为此,提出一种基于改进GMM的运动目标检测方法。利用不匹配像素消除光照影响,使用改进的GMM提取背景图像。通过差分当前帧与背景图像获得二值差分图像,从该差分图像中获取运动目标。实验结果表明,该方法能适应光照变化,提高检测的准确性和鲁棒性。     

结合深度学习和超像元的高分遥感影像变化检测

目的 随着遥感影像空间分辨率的提升,相同地物的空间纹理表现形式差异变大,地物特征更加复杂多样,传统的变化检测方法已很难满足需求。为提高高分辨率遥感影像的变化检测精度,尤其对相同地物中纹理差异较大的区域做出有效判别,提出结合深度学习和超像元分割的高分辨率遥感影像变化检测方法。方法 将有限带标签数据分割成切片作训练样本,按照样本形式设计一个多切片尺度特征融合网络并对其训练,获得测试图像的初步变化检测结果;利用超像元分割算法将测试图像分割成许多无重叠的同质性区域,并将分割结果与前述检测结果叠合,得到带分割标记的变化检测结果;用举手表决算法统计带分割标记的变化检测结果中超像元的变化状况,得到最终变化检测结果。结果 在变化检测实验结果中,本文提出的多切片尺度特征融合卷积网络模型在广东数据集和香港数据集上,优于单一切片尺度下卷积神经网络模型,并且结合超像元的多切片尺度特征融合卷积网络模型得到的Kappa系数分别达到80%和82%,比相应的非超像元算法分别提高了6%和8%,在两个测试集上表现均优于长短时记忆网络、深度置信网络等对比算法。结论 本文提出的卷积神经网络变化检测方法可以充分学习切片的空间信息和其他有效特征,避免过拟合现象;多层尺度切片特征融合的方法优于单一切片尺度训练神经网络的方法;结合深度学习和超像元分割算法,检测单元实现了由切片到超像元的转变,能对同物异谱的区域做出有效判决,有利于提升变化检测精度。     

Histogram thresholding for unsupervised change detection of remote sensing images

The change-detection problem can be viewed as an unsupervised classification problem with two classes corresponding to changed and unchanged areas. Image differencing is a widely used approach to change detection. It is based on the idea of generating a difference image that represents the modulus of the spectral change vectors associated with each pixel in the study area. To separate out the changed and unchanged classes in the difference image automatically, any unsupervised technique can be used. Thresholding is one of the cheapest techniques among them. However, in thresholding approaches, selection of the best threshold value is not a trivial task. In this work, several non-fuzzy and fuzzy histogram thresholding techniques are investigated and compared for the change-detection problem. Experimental results, carried out on different multitemporal remote sensing images (acquired before and after an event), are used to assess the effectiveness of each of the thresholding techniques. Among all the thresholding techniques investigated here, Liu's fuzzy entropy followed by Kapur's entropy are found to be the most robust techniques.     

Unsupervised change detection of VHR remote sensing images based on multi-resolution Markov Random Field in wavelet domain

This paper proposes an unsupervised change detection method for very-high-resolution (VHR) remote sensing images based on multi-resolution Markov random field (MRF) model in wavelet domain. Firstly, the wavelet transform is performed on the difference image achieved by the change vector analysis (CVA) method, and the wavelet coefficients at each scale are obtained. Then, MRF model is constructed based on the wavelet coefficients. The wavelet high-frequency coefficients establish a feature field model that describes the feature attributes of each pixel location at each scale. The initial change map (changed and unchanged) at the coarse scale are generated through applying the k-means method to the wavelet low-frequency coefficients, and a label field model describing the region of the variation results is established. The label and feature field, at the same scale, got the optimized change map under the Bayesian criterion. Finally, the results of the low-resolution scale change map are directly projected as the adjacent higher-scale initial change map. The more accurate change map is obtained successively from the coarse scale to the original resolution scale, and the detection result of the original resolution is obtained at last. Experiments on Quick Bird, SPOT-5, and IKONOS optical images have demonstrated the effectiveness of the proposed method. The experimental results show that the method has better regional consistency and strong robustness.     

Incorporating global-local a Priori knowledge into expectation-maximization for SAR image change detection

Speckle is one of the inevitable obstacles related to synthetic aperture radar (SAR) image change detection; it increases the overlap between changed and unchanged pixels in the histogram of a difference image. This makes the selection of a statistic model more difficult for describing opposite classes. To address this issue, this article developed an unsupervised change-detection approach for multitemporal SAR images that specifies a priori knowledge about the spatial characteristics of the classes through Dempster-Shafer evidence theory and embeds it into the Expectation-Maximization (EM) iteration process. It is based on the consideration that each pixel in the difference image is unique due to its neighbourhood, although some of them may have the same pixel value. Thus, under the hypothesis that local and global a priori knowledge are independent sources, a global-local a priori model is developed through Dempster-Shafer evidence theory. The EM algorithm allows one to estimate the statistical parameters of the opposite classes associated with this a priori model. As a consequence, the change-detection result can be obtained within the framework of Bayes. Visual and quantitative results obtained on real multitemporal SAR image data sets confirm the effectiveness of the proposed method compared with state-of-the-art ones for SAR image change detection.     

Accelerated genetic algorithm based on search-space decomposition for change detection in remote sensing images

Detecting change areas among two or more remote sensing images is a key technique in remote sensing. It usually consists of generating and analyzing a difference image thus to produce a change map. Analyzing the difference image to obtain the change map is essentially a binary classification problem, and can be solved by optimization algorithms. This paper proposes an accelerated genetic algorithm based on search-space decomposition (SD-aGA) for change detection in remote sensing images. Firstly, the BM3D algorithm is used to preprocess the remote sensing image to enhance useful information and suppress noises. The difference image is then obtained using the logarithmic ratio method. Secondly, after saliency detection, fuzzy c-means algorithm is conducted on the salient region detected in the difference image to identify the changed, unchanged and undetermined pixels. Only those undetermined pixels are considered by the optimization algorithm, which reduces the search space significantly. Inspired by the idea of the divide-and-conquer strategy, the difference image is decomposed into sub-blocks with a method similar to down-sampling, where only those undetermined pixels are analyzed and optimized by SD-aGA in parallel. The category labels of the undetermined pixels in each sub-block are optimized according to an improved objective function with neighborhood information. Finally the decision results of the category labels of all the pixels in the sub-blocks are remapped to their original positions in the difference image and then merged globally. Decision fusion is conducted on each pixel based on the decision results in the local neighborhood to produce the final change map. The proposed method is tested on six diverse remote sensing image benchmark datasets and compared against six state-of-the-art methods. Segmentations on the synthetic image and natural image corrupted by different noise are also carried out for comparison. Results demonstrate the excellent performance of the proposed SD-aGA on handling noises and detecting the changed areas accurately. In particular, compared with the traditional genetic algorithm, SD-aGA can obtain a much higher degree of detection accuracy with much less computational time.     

基于PCA和GMM的图像分类算法

讨论了目标图像类和非目标图像类的分类方法.按统计学原理,如果图像类属于目标图像类,则提取图像中目标图像的特征,否则提取整幅图像的底层特征,基于主分量分析（PCA）的图像特征降维方法和高斯混合模型（GMM）分类器,提出了一种图像分类算法,该算法在标准的Corel图像库上进行了测试,并与其它基于GMM的方法进行了比较,实验结果表明了提出算法的有效性.     

基于Laplacian Eigenmap的图像变化检测虚警优化技术

对点目标的图像变化检测,现有的变化检测技术结果往往存在着虚警过大的问题。通过深入分析多个传统的变化检测方法的特点,利用各方法的互补性,提出了利用Laplacian Eigenmap对多个方法检测结果进行降维分类的优化技术。首先把各个方法对某个像素的检测结果用向量的形式进行表示,然后利用Laplacian Eigenmap对整个图像的数据流形在低维空间展开,最后利用模糊分类进行分类。该技术有两个优势：（1）在保证现有较高检测率的同时,大大降低了结果的虚警率;（2）它极大地降低了在传统方法中由于人为阈值取舍带来的偏差风险。但该技术的不足之处是增加了计算量。     

结合结构相似度的自适应多尺度SAR图像变化检测

目的 结合高斯核函数特有的性质,提出一种基于结构相似度的自适应多尺度SAR图像变化检测算法。方法 本文提出的算法包括差异图像获取、高斯多尺度分解、基于结构相似性的最优尺度选择、特征矢量构造以及模糊C均值分类。首先,通过对多时相SAR图像进行对数比运算获取差异图像,然后,利用基于图像的结构相似度估计高斯多尺度变换的最优尺度,继而在该最优尺度参数下逐像素构建变化检测特征矢量,最后通过模糊C均值聚类方法实现变化像素与未变化像素的分离,生成最终的变化检测结果图。结果 在两组真实的SAR图像数据上测试本文算法,正确检测率分别达到0.9952和0.9623,Kappa系数分别为0.8200和0.8540,相比传统算法有了较大的提高。结论 本文算法充分利用了尺度信息,对噪声的鲁棒性有所提高。实测SAR数据的实验结果表明,本文算法可以智能获取最优分解尺度,显著提高了SAR图像变化检测性能。     

融合双特征图信息的图像显著性检测方法

目的 图像的显著性检测是将图像中最重要的、包含丰富信息的区域标记出来,并应用到图像分割、图像压缩、图像检索、目标识别等重要领域。针对现有研究方法显著性目标检测结果不完整以及单一依靠颜色差异检测方法的局限性,提出一种综合图像底层颜色对比特征图和图像颜色空间分布特征图的显著性检测方法,能够有效而完整地检测出图像中的显著性区域。方法 本文方法结合了SLIC超像素分割和K-means聚类算法进行图像特征的提取。首先,对图像进行SLIC（simple linear iterative clustering）分割,根据像素块之间的颜色差异求取颜色对比特征图;其次,按照颜色特征对图像进行K-means聚类,依据空间分布紧凑性和颜色分布统一性计算每个类的初步颜色空间分布特征。由于聚类结果中不包含空间信息,本文将聚类后的结果映射到超像素分割的像素块上,进一步优化颜色空间分布图;最后,通过融合颜色对比显著图和图像颜色空间分布特征图得到最终的显著图。结果 针对公开的图像测试数据库MSRA-1000,本文方法与当前几种流行的显著性检测算法进行了对比实验,实验结果表明,本文方法得到的显著性区域更准确、更完整。结论 本文提出了一种简单有效的显著性检测方法,结合颜色对比特征图和图像颜色空间分布特征图可以准确的检测出显著性区域。该结果可用于目标检测等实际问题,但该方法存在一定的不足,对于背景色彩过于丰富且与特征区域有近似颜色的图像,该方法得到的结果有待改进。今后对此算法的优化更加侧重于通用性。     

Difference representation learning using stacked restricted Boltzmann machines for change detection in SAR images

In this paper, we establish a deep neural network using stacked Restricted Boltzmann Machines (RBMs) to analyze the difference images and detect changes between multitemporal synthetic aperture radar (SAR) images. Given the two multitemporal images, a difference image which shows difference degrees between corresponding pixels is generated. Then, RBMs are stacked to form a deep hierarchical neural network to learn to analyze the difference image and recognize the changed pixels and unchanged pixels. The learning process includes unsupervised layer-wise feature learning and supervised fine-tuning of network parameters. Unsupervised learning aims to learn the representation of the difference image. Supervised fine-tuning aims to learn to classify the changed and unchanged pixels. The network can learn from datasets that have few labeled data. The labeled data can be selected from the results obtained by other methods because there is no prior information in image change detection. The system learns to detect the changes instead of recognizing the changes by fixed equations as in traditional change detection algorithms. We test the network with real synthetic aperture radar datasets and the labeled samples are extracted from the results obtained, respectively, by several methods, including a thresholding method, a level set method and two clustering methods. The results achieved by the trained network outperform that of other methods.     

融入邻域作用的高斯混合分割模型及简化求解

目的 基于高斯混合模型（GMM）的图像分割方法易受噪声影响,为此采用马尔可夫随机场（MRF）将像素邻域关系引入GMM,提高算法抗噪性。针对融入邻域作用的高斯混合分割模型结构复杂、参数估计困难,难以获得全局最优分割解等问题,提出一种融入邻域作用的高斯混合分割模型及其简化求解方法。方法 首先,构建融入邻域作用的GMM。为了提高GMM的抗噪性,采用MRF建模混合模型权重系数的先验分布。然后,利用贝叶斯理论建立图像分割模型,即品质函数;由于品质函数中参数较多（包括权重系数,均值,协方差）、函数结构复杂,导致参数求解困难。因此,将品质函数中的均值和协方差定义为权重系数的函数,由此简化模型结构并方便其求解;虽然品质函数中仅包含参数权重系数,但结构比较复杂,难以求得参数的解析式。最后,采用非线性共轭梯度法（CGM）求解参数,该方法仅需利用品质函数值和参数梯度值,降低了参数求解的复杂性,并且收敛快,可以得到全局最优解。结果 为了有效而准确地验证提出的分割方法,分别采用本文算法和对比算法对合成图像和高分辨率遥感图像进行分割实验,并定性和定量地评价和分析了实验结果。实验结果表明本文方法的有效抗噪性,并得到很好的分割结果。从参数估计结果可以看出,本文算法有效简化了模型参数,并获得全局最优解。结论 提出一种融入邻域作用的高斯混合分割模型及其简化求解方法,实验结果表明,本文算法提高了算法的抗噪性,有效地简化了模型参数,并得到全局最优参数解。本文算法对具有噪声的高分辨率遥感影像广泛适用。     

光线变化下的视频图像分割

视频图像分割是视频目标定位和识别的基础,如果背景中光线变化,那么将会给分割带来很大的影响。文中利用贝叶斯学习方法进行视频图像分割,在每个象素点处对不断变化的背景建模,计算每个象素点处的颜色直方图,用这些直方图来表示该象素点处特征向量的概率分布,然后用贝叶斯学习方法来判断,在光线缓慢或者突然变化的时候,每个象素点是属于前景还是属于背景。     

一种新的结合非下采样Contourlet域融合和参数化内核图割的SAR图像无监督水灾变化检测

目的：基于非下采样Contourlet变换(NSCT)融合策略可以有效地抑制背景信息增强变化区域的信息。但是融合后图像具有复杂的统计特征,传统的基于统计特征的变化检测难以实现。基于参数化内核图割的遥感图像分割不受统计特征的限制。为此提出了一种基于NSCT融合和参数化内核图割的SAR图像无监督水灾变化检测新算法。方法：将均值比差异图像和对数比差异图像采用基于NSCT的融合算法进行融合,将融合后的差异图像采用参数化内核图割算法进行前景/背景的分割,得到最终的变化检测结果。结果：融合后的差异图像利用前两种差异图像的互补信息提高了变化检测精度。算法不受统计模型限制,不需要先验知识,适用性强。结论：实验结果表明,本文算法的检测精度优于传统的变化检测方法。     

结合孪生网络和像素配对的高光谱图像异常检测

目的 高光谱遥感中,通常利用像素的光谱特征来区分背景地物和异常目标,即通过二者之间的光谱差异来寻找图像中的异常像元。但传统的异常检测算法并未有效挖掘光谱的深层特征,高光谱图像中丰富的光谱信息没有被充分利用。针对这一问题,本文提出结合孪生神经网络和像素配对策略的高光谱图像异常检测方法,利用深度学习技术提取高光谱图像的深层非线性特征,提高异常检测精度。方法 采用像素配对的思想构建训练样本,与原始数据集相比,配对得到的新数据集数量呈指数增长,从而满足深度网络对数据集数量的需求。搭建含有特征提取模块和特征处理模块的孪生网络模型,其中,特征处理模块中的卷积层可以专注于提取像素对之间的差异特征,随后利用新的训练像素对数据集进行训练,并将训练好的分类模型固定参数,迁移至检测过程。用滑动双窗口策略对测试集进行配对处理,将测试像素对数据集送入网络模型,得到每个像素相较于周围背景像素的差异性分数,从而识别测试场景中的异常地物。结果 在异常检测的实验结果中,本文提出的孪生网络模型在San Diego数据集的两幅场景和ABU-Airport数据集的一幅场景上,得到的AUC （area under the curve）值分别为0.993 51、0.981 21和0.984 38,在3个测试集上的表现较传统方法和基于卷积神经网络的异常检测算法具有明显优势。结论 本文方法可以提取输入像素对的深层光谱特征,并根据其特征的差异性,让网络学习到二者的区分度,从而更好地赋予待测像素相对于周围背景的异常分数。本文方法相对于卷积神经网络的异常检测方法可以有效地降低虚警,与传统方法相比能够更加明显地突出异常目标,提高了检测率,同时也具有较强的鲁棒性。     

基于贝叶斯学习的视频图像分割

如果背景中光线变化,那么视频图像分割将会变得比较困难。为了对光线变化的图像进行顺利侵害,提出了一种利用贝叶斯学习方法来进行视频图像分割的算法,即先在每个像素点处对不断变化的背景建模,同时计算每个像素点处的颜色直方图,再用这些直方图来表示该像素点处特征向量的概率分布,然后用贝叶斯学习方法来进行判断,以确定在光线缓慢或者突然变化的时候,每个像素点是属于前景还是属于背景。     

A genetic expectation-maximization method for unsupervised change detection in multitemporal SAR imagery

In this paper, we present a novel, automatic and unsupervised change-detection approach to the analysis of single-channel single-polarization multitemporal SAR images. The statistical parameters of the changed and unchanged classes, which are assumed to follow a generalized Gaussian (GG) distribution in the analysed log-ratio image, are explicitly estimated by the expectation-maximization (EM) algorithm initialized with a robust strategy based on genetic algorithms (GAs). In addition, the proposed approach integrates two further processing capabilities. The first one intends to cope with the problem of the automatic detection of multiple changes in the scene. This is carried out by modelling the log-ratio image histogram with a multimodal GG mixture whose number of components is estimated basing on the Bayesian information criterion (BIC). The second processing capability allows exploitation of spatial contextual information in the change detection process through a Markovian formulation. Results obtained on both simulated and real data are reported and discussed.