A modified mean filter for improving the classification performance of very high-resolution remote-sensing imagery

Very high resolution (VHR) remote-sensing imagery can reveal ground objects in great detail, depicting the colour, shape, size and structure of the objects. However, VHR also leads to a large amount of noise in the spectra, which may reduce the reliability of the classification result. This article presents an extension of the mean filter (MF), which is named ‘modified mean filter (MMF)’, for smoothing the noise of VHR imagery. First, the MMF is a shape-adaptive filter that is constructed by gradually detecting the spectral similarity between a kernel-anchored pixel and its contextual pixels through an extension detector with eight neighbouring pixels. Then, because pixels of an objective are usually homogeneous with spatial continuity, the pixels located at the hollow area within an extended region are rectified to enhance the smoothing effect. Finally, the spectral value of the kernel-anchored pixel is determined by the mean of the group of pixels within the adaptive region. Despite the proposed filter is a simple extension of MF, it has an advantage in preserving the edge between different classes, and smoothing the noise of intra-class. The MMF approach is investigated through comparing with the classification of VHR images based on filter processing, including the traditional mean filter (MF), the median filter (MedF) and the recursive filter (RF) which has been proposed for image classification in Kang, Li, and Benediktsson (2014). The experimental results obtained by considering two VHR images show the effectiveness of the proposed of MMF, which improves the performance of the classification and implies more potential applications.     

Unsupervised change detection using spectral features and a texture difference measure for VHR remote-sensing images

This article proposes an unsupervised change-detection method using spectral and texture information for very-high-resolution (VHR) remote-sensing images. First, a new local-similarity-based texture difference measure (LSTDM) is defined using a grey-level co-occurrence matrix. A mathematical analysis shows that LSTDM is robust with respect to noise and spectral similarity. Second, the difference image is generated by integrating the spectral and texture features. Then, the unsupervised change-detection problem in VHR remote-sensing images is formulated as minimizing an energy function related with changed and unchanged classes in the difference image. A modified expectation-maximization-based active contour model (EMCVM) is applied to the difference image to separate the changed and unchanged regions. Finally, two different experiments are performed with SPOT-5 images and compared with state-of-the-art unsupervised change-detection methods to evaluate the effectiveness of the proposed method. The results indicate that the proposed method can sufficiently increase the robustness with respect to noise and spectral similarity and obtain the highest accuracy among the methods addressed in this article.     

Using random walker for knowledge transfer in classifying multi-temporal VHR images

Acquiring land cover types from very high resolution (VHR) images is of great significance to many applications and has been intensively studied for many years. The difficulties in image classification and the high frequencies of remote sensing image acquisition make it urgent to develop efficient knowledge transfer approaches for understanding multi-temporal VHR images. This letter proposed a knowledge transfer approach that uses the label information of the existing VHR images to classify multi-temporal images. The approach was implemented in three steps: object-based change detection, knowledge transfer of label information, and random walker (RW) classification. The proposed approach was tested by two datasets with each having two temporal images acquired on the same geographical areas. The experimental results showed that the proposed approach outperformed the support vector machine (SVM) algorithm in classifying multi-temporal images and can reduce the influence of spectral confusions on image classification.     

Interactive image segmentation using probabilistic hypergraphs

This paper introduces a novel interactive framework for segmenting images using probabilistic hypergraphs which model the spatial and appearance relations among image pixels. The probabilistic hypergraph provides us a means to pose image segmentation as a machine learning problem. In particular, we assume that a small set of pixels, which are referred to as seed pixels, are labeled as the object and background. The seed pixels are used to estimate the labels of the unlabeled pixels by learning on a hypergraph via minimizing a quadratic smoothness term formed by a hypergraph Laplacian matrix subject to the known label constraints. We derive a natural probabilistic interpretation of this smoothness term, and provide a detailed discussion on the relation of our method to other hypergraph and graph based learning methods. We also present a front-to-end image segmentation system based on the proposed method, which is shown to achieve promising quantitative and qualitative results on the commonly used GrabCut dataset.     

Fuzzy clustering algorithms for unsupervised change detection in remote sensing images

In this paper, we propose a context-sensitive technique for unsupervised change detection in multitemporal remote sensing images. The technique is based on fuzzy clustering approach and takes care of spatial correlation between neighboring pixels of the difference image produced by comparing two images acquired on the same geographical area at different times. Since the ranges of pixel values of the difference image belonging to the two clusters (changed and unchanged) generally have overlap, fuzzy clustering techniques seem to be an appropriate and realistic choice to identify them (as we already know from pattern recognition literatures that fuzzy set can handle this type of situation very well). Two fuzzy clustering algorithms, namely fuzzy c-means (FCM) and Gustafson-Kessel clustering (GKC) algorithms have been used for this task in the proposed work. For clustering purpose various image features are extracted using the neighborhood information of pixels. Hybridization of FCM and GKC with two other optimization techniques, genetic algorithm (GA) and simulated annealing (SA), is made to further enhance the performance. To show the effectiveness of the proposed technique, experiments are conducted on two multispectral and multitemporal remote sensing images. A fuzzy cluster validity index (Xie-Beni) is used to quantitatively evaluate the performance. Results are compared with those of existing Markov random field (MRF) and neural network based algorithms and found to be superior. The proposed technique is less time consuming and unlike MRF does not require any a priori knowledge of distributions of changed and unchanged pixels.     

Change detection of multispectral remote-sensing images using stationary wavelet transforms and integrated active contours

This article presents a change-detection approach for multispectral remote-sensing images. In our approach, we first exploit a wavelet-based, multi-resolution representation of the difference image. For each resolution scale, the multispectral difference image representation is considered as a 2-D Riemannian manifold embedded into a Riemannian manifold with a higher dimensionality. The integrated active contour (IAC) model is then applied to the multiband difference image representation to obtain a change-detection map at a given scale. In order to select a reasonable scale for each pixel, a measurement of regional homogeneity is defined by comparing the determinant of the metric with the average value of the metric’s determinant. For a single pixel, the final change-detection result is generated by selecting the change map on a reasonable scale. Experimental results obtained on multispectral remote-sensing images confirm the effectiveness of the proposed approach, although the time consumption of the approach is somewhat high. Our experiment achieved total error rates of 3.41%, 1.00%, and 1.95% for three data sets, which are comparable to other prevalent algorithms.     

Root Mean Square filter for noisy images based on hyper graph model

In this paper, we propose a noise removal algorithm for digital images. This algorithm is based on hypergraph model of image, which enables us to distinguish noisy pixels in the image from the noise-free ones. Hence, our algorithm obviates the need for denoising all the pixels, thereby preserving as much image details as possible. The identified noisy pixels are denoised through Root Mean Square (RMS) approximation. The performance of our algorithm, based on peak-signal-to-noise-ratio (PSNR) and mean-absolute-error (MAE), was studied on various benchmark images, and found to be superior to that of other traditional filters and other hypergraph based denoising algorithms.     

An approach based on discrete wavelet transform to unsupervised change detection in multispectral images

Change vector analysis (CVA) and Spectral Angle Mapper (SAM) are widely used for change detection in multitemporal multispectral images. CVA and SAM describe the difference from the perspective of vector magnitude and spectral angle, respectively. It has been proved that three change categories may occur in a changed pixel; however, CVA or SAM alone can only detect two of the three change categories properly. Hence, we propose a novel approach integrating the advantages of them to acquire a better change map. This approach, based on discrete wavelet transform (ABDWT, i.e. approach based on discrete wavelet transform), obtains two difference images by using CVA and SAM, and then yields a novel difference image by fusing them in the coefficients domains of discrete wavelet transform. Experimental results from a simulated and two real data sets validate the effectiveness of the proposed approach. In the first real data set, the proposed approach can identify 14,916 changed pixels while the best result of other methods is 14,806. In the second real data set, the proposed approach detects 3203 changed pixels, while the maximum of other methods is 3189.     

基于超图切割的半监督学习和聚类算法

本文针对超图切割上的半监督学习和聚类算法进行了研究;首先,通过对超图切割和超边展开法及其切割函数的讨论,引入了超图上的总变异作为超图切割的洛瓦兹扩展,并在此基础上提出了一组正则化函数,它对应于图上的拉普拉斯型正则化;然后,基于正则化函数族提出了半监督学习方法,并基于平衡超图切割提出了谱聚类方法;为了求解这两个学习问题,将它们转化为求解凸优化问题,并为此提出了一种主要组成部分为近端映射的可扩展算法,从而实现半监督学习和聚类;仿真实验结果表明,本文提出的基于超图切割实现的半监督学习和聚类方法相比于经典的超边展开法和其他图切割方法有更好的标准偏差和聚类误差性能。     

Spatial–spectral locality constrained elastic net hypergraph for hyperspectral image clustering

Hypergraph is an effective method used to represent the contextual correlation within hyperspectral imagery for clustering. Nevertheless, how to discover the closely correlated samples to form hyperedges is the key issue for constructing an informative hypergraph. In this article, a new spatial–spectral locality constrained elastic net hypergraph learning model is proposed for hyperspectral image clustering (i.e. unsupervised classification). In order to utilize the spatial–spectral correlation among the pixels in hyperspectral images, first, we construct a locality-constrained dictionary by selecting K relevant pixels within a spatial neighbourhood, which activates the most correlated atoms and suppresses the uncorrelated ones. Second, each pixel is represented as a linear combination of the atoms in the dictionary under the elastic net regularization. Third, based on the obtained representations, the pixels and their most related pixels are linked as hyperedges, which can effectively capture high–order relationships among the pixels. Finally, a hypergraph Laplacian matrix is built for unsupervised learning. Experiments have been conducted on two widely used hyperspectral images, and the results show that the proposed method can achieve a superior clustering performance when compared to state-of-the-art methods.     

An image fusion method based on object-oriented classification

In image fusion of different spatial resolution multispectral (MS) and panchromatic (PAN) images, a spectrally mixed MS pixel superimposes multiple mixed PAN pixels and multiple pure PAN pixels. This verifies that with increased spatial resolution in imaging, a low spatial resolution spectrally mixed subpixel may be unmixed to be a pure pixel. However, spectral unmixing of mixed MS subpixels is rarely considered in current remote-sensing image fusion methods, resulting in blurred fused images. In the image fusion method proposed in this article, such spectral unmixing is realized. In this method, the MS and PAN images are jointly segmented into image objects, image objects are classified to obtain a classification map of the PAN image and each MS subpixel is fused to be a pixel matching the class of the corresponding PAN pixel. Tested on spatially degraded IKONOS MS and PAN images with a significant spatial resolution ratio of 8:1, the fusion method offered fused images with high spectral quality and deblurred visualization.     

Probabilistic hypergraph based hash codes for social image search

With the rapid development of the Internet, recent years have seen the explosive growth of social media. This brings great challenges in performing efficient and accurate image retrieval on a large scale. Recent work shows that using hashing methods to embed high-dimensional image features and tag information into Hamming space provides a powerful way to index large collections of social images. By learning hash codes through a spectral graph partitioning algorithm, spectral hashing（SH） has shown promising performance among various hashing approaches. However, it is incomplete to model the relations among images only by pairwise simple graphs which ignore the relationship in a higher order. In this paper, we utilize a probabilistic hypergraph model to learn hash codes for social image retrieval. A probabilistic hypergraph model offers a higher order repre-sentation among social images by connecting more than two images in one hyperedge. Unlike a normal hypergraph model, a probabilistic hypergraph model considers not only the grouping information, but also the similarities between vertices in hy-peredges. Experiments on Flickr image datasets verify the performance of our proposed approach.     

Combining Active and Semi-Supervised Learning for Video Compression

Video compression algorithms manipulate video signals to dramatically reduce the storage and bandwidth required while maximizing perceived video quality. Typical video compression methods include discrete cosine transform, vector quantization, fractal compression, and discrete wavelet transform. Recently, a machine learning based approach has been proposed which converts the color images (frames) to gray scale images (frames) and the color information for only a few representative pixels is kept. A learning model is then trained to predict the color values for the gray scale pixels across frames. Selecting the most representative pixels is essentially an active learning problem, while colorization is a semi-supervised learning problem. In this paper, we propose to combine active and semi-supervised learning for video compression. The basic idea is to minimize the size of the covariance matrix of the regularized least squares estimates, in which the regression model assumes that each pixel can be reconstructed by the other pixels with similar spatial location and intensity value. The experimental results demonstrate the effectiveness of the proposed approach for video compression.     

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.     

Improving spatiotemporal reflectance fusion using image inpainting and steering kernel regression techniques

A novel spatiotemporal reflectance fusion method integrating image inpainting and steering kernel regression fusion model (ISKRFM) is proposed to improve the fusion accuracy for remote-sensing images with different temporal and spatial characteristics in this article. This method first detects the land-cover changed regions and then fills them with unchanged similar pixels by an exemplar-based inpainting technique. Furthermore, a steering kernel regression (SKR) is used to adaptively determine the weightings of local neighbouring pixels to predict high spatial resolution image. Accordingly, the main contributions of this method are twofold. One is to address the land-cover change issues in the spatiotemporal fusion, and the other is to establish an adaptive weighting assignment according to the pixel locations and the radiometric properties of the local neighbours to account for the effect of neighbouring pixels. To validate the proposed method, two actual Enhanced Thematic Mapper Plus (ETM+) and Moderate Resolution Imaging Spectroradiometer (MODIS) acquisitions at southeast China were implemented and compared with the baseline spatial and temporal adaptive reflectance fusion model (STARFM). The experimental results demonstrate that addressing the land-cover changes in spatiotemporal fusion has positive effects on the fused image, and the proposed ISKRFM method significantly outperforms STARFM in terms of both visual and quantitative measurements.     

Object-based change detection from satellite imagery by segmentation optimization and multi-features fusion

This article presents a novel object-based change detection (OBCD) approach in high-resolution remote-sensing images by means of combining segmentation optimization and multi-features fusion. In the segmentation optimization, objects with optimized boundaries and proper sizes are generated by object intersection and merging (OIM) processes, which ensures the accurate information extraction from image objects. Within multi-features fusion and change analysis, the Dempster and Shafer (D-S) evidence theory and the Expectation-Maximization (EM) algorithm are implemented, which effectively utilize multidimensional features besides avoiding the selection of an appropriate change threshold. The main advantages of our proposed method lie in the improvement of object boundary and the fuzzy fusion of multi-features information. The proposed approach is evaluated using two different high-resolution remote-sensing data sets, and the qualitative and quantitative analyses of the results demonstrate the effectiveness of the proposed approach.     

Pointwise SAR image change detection based on stereograph model with multiple-span neighbourhood information

In this article, a novel pointwise approach is proposed for change detection in bi-temporal synthetic aperture radar (SAR) images using stereograph model. Due to the fact that SAR image suffers from the speckle noise, a pointwise approach based on a set of characteristic points only, not on the whole pixels, seems to be more efficient. Moreover, the correlations of neighbourhood points which have different locations in bi-temporal SAR images should be studied to repress the speckle in change detection. Therefore, the stereograph model, which extends the graph model to three-dimensional space, is designed to connect the local maximum pixels on bi-temporal SAR images and can be used to capture the multiple-span neighbourhood information from the edges. Furthermore, a specialized change measure function is presented to quantify the neighbourhood information from stereograph model, and thus, a novel nondense difference image (NDI) is generated. Finally, a traditional classification method is used to analyse the NDI into changed class and unchanged class. Experiments on real SAR images show that the proposed NDI can improve separability between changed and unchanged areas, and the final results possess high accuracy and strong noise immunity for change detection tasks with noise-contaminated SAR images.     

视角和光照显著变化时的变化检测方法研究

探讨使用计算机视觉的最新方法来解决基于两幅高空间分辨率光学遥感图像的城市变化检测问题. 基本原理是通过提取聚类出现的变化直线段群来提取城市变化, 重点研究了拍摄视角和光照条件显著变化时的几个主要问题. 提出了一种基于多种类型图像特征的匹配方法来提取无变化建筑的顶部区域, 结合几何约束引入了变化盲区的概念以处理高层建筑在不同视角和光照下的图像不同现象. 使用真实遥感图像进行实验, 在视角和光照显著变化时仍可取得满意的变化检测结果.     

Object-based land-cover change detection applied to Brazilian seasonal savannahs using geostatistical features

A new method for remote-sensing land-use/land-cover (LULC) change detection is proposed to eliminate the effects of forest phenology on classification results. This method is insensitive to spectral changes caused by vegetation seasonality and uses an object-based approach to extract geostatistical features from bitemporal Landsat TM (Thematic Mapper) images. We first create image objects by multiresolution segmentation to extract geostatistical features (semivariogram parameters and indices) and spectral information (average values) from NDVI (normalized difference vegetation index), acquired in the wet and dry seasons, as input data to train a Support Vector Machine algorithm. We also used the image difference traditional change-detection method to validate the effectiveness of the proposed method. We used two classes: (1) LULC change class and (2) seasonal change class. Using the most geostatistical features, the change detection results are considerably improved compared with the spectral features and image differencing technique. The highest accuracy was achieved by the sill (σ² overall variability) semivariogram parameter (95%) and the AFM (area first lag–first maximum) semivariogram index (88.33%), which were not affected by vegetation seasonality. The results indicate that the geostatistical context makes possible the use of bitemporal NDVI images to address the challenge of accurately detecting LULC changes in Brazilian seasonal savannahs, disregarding changes caused by phenological differences, without using a dense time series of remote-sensing images. The challenge of extracting accurate semivariogram curves from objects of long and narrow shapes requires further study, along with the relationship between the scale of segmentation and image spatial resolution, including the type of change and the initial land-cover class.     

Automatic normalization of satellite images using unchanged pixels within urban areas

Radiometric normalization is often required in remote sensing image analysis particularly in land change analysis. The normalization minimizes different imaging condition effects in analysis and rectifies radiometry of images in such a way as if they have been acquired at the same imaging conditions. Relative radiometric normalization which is normally applied in image preprocessing stage does not remove all unwanted effects. In this paper, an automatic normalization method has been developed based on regression applied on unchanged pixels within urban areas. The proposed method is based on efficient selection of unchanged pixels through image difference histogram modeling using available spectral bands and calculation of relevant coefficients for dark, gray and bright pixels in each band. The coefficients are applied to produce the normalized image. The idea has been implemented on two TM image datasets. The capability of the approach in taking into account the imaging condition differences and effectively excluding real land change pixels from the normalization process has shown better performance in the evaluation stage.