相似性判别原理在影像变化检测中的应用

针对传统基于Bayes影像变化监测方法中分类的硬划分及对辐射畸变敏感的问题,首先,采用集合相似性测度动态更新变化和未变化两类像元模糊子集,解决变化检测分类过程中硬划分的问题;然后,采用像元相似性测度来判断变化类像元集合和未变化类像元集合中的每个像元是否真正发生了变化,剔除伪变化像元,克服对影像线性辐射畸变敏感的问题,并检测出漏检像元,最终实现影像变化区域的提取。实验结果表明,基于相似性测度的影像变化检测方法可实现变化区域自动化提取,具有较好的应用价值。     

A modified stochastic neighbor embedding for multi-feature dimension reduction of remote sensing images

In automated remote sensing based image analysis, it is important to consider the multiple features of a certain pixel, such as the spectral signature, morphological property, and shape feature, in both the spatial and spectral domains, to improve the classification accuracy. Therefore, it is essential to consider the complementary properties of the different features and combine them in order to obtain an accurate classification rate. In this paper, we introduce a modified stochastic neighbor embedding (MSNE) algorithm for multiple features dimension reduction (DR) under a probability preserving projection framework. For each feature, a probability distribution is constructed based on t-distributed stochastic neighbor embedding (t-SNE), and we then alternately solve t-SNE and learn the optimal combination coefficients for different features in the proposed multiple features DR optimization. Compared with conventional remote sensing image DR strategies, the suggested algorithm utilizes both the spatial and spectral features of a pixel to achieve a physically meaningful low-dimensional feature representation for the subsequent classification, by automatically learning a combination coefficient for each feature. The classification results using hyperspectral remote sensing images (HSI) show that MSNE can effectively improve RS image classification performance.     

Using high resolution QuickBird imagery for crop identification and area estimation

Imagery from recently launched high spatial resolution satellite sensors offers new opportunities for crop assessment and monitoring. A 2.8-m multispectral QuickBird image covering an intensively cropped area in south Texas was evaluated for crop identification and area estimation. Three reduced-resolution images with pixel sizes of 11.2 m, 19.6 m, and 30.8 m were also generated from the original image to simulate coarser resolution imagery from other satellite systems. Supervised classification techniques were used to classify the original image and the three aggregated images into five crop classes (grain sorghum, cotton, citrus, sugarcane, and melons) and five non-crop cover types (mixed herbaceous species, mixed brush, water bodies, wet areas, and dry soil/roads). The five non-crop classes in the 10-category classification maps were then merged as one class. The classification maps were filtered to remove the small inclusions of other classes within the dominant class. For accuracy assessment of the classification maps, crop fields were ground verified and field boundaries were digitized from the original image to determine reference field areas for the five crops. Overall accuracy for the unfiltered 2.8-m, 11.2-m, 19.6-m, and 30.8-m classification maps were 71.4, 76.9, 77.1, and 78.0%, respectively, while overall accuracy for the respective filtered classification maps were 83.6, 82.3, 79.8, and 78.5%. Although increase in pixel size improved overall accuracy for the unfiltered classification maps, the filtered 2.8-m classification map provided the best overall accuracy. Percentage area estimates based on the filtered 2.8-m classification map (34.3, 16.4, 2.3, 2.2, 8.0, and 36.8% for grain sorghum, cotton, citrus, sugarcane, melons, and non-crop, respectively) agreed well with estimates from the digitized polygon map (35.0, 17.9, 2.4, 2.1, 8.0, and 34.6% for the respective categories). These results indicate that QuickBird imagery can be a useful data source for identifying crop types and estimating crop areas.     

A spatial–temporal Hopfield neural network approach for super-resolution land cover mapping with multi-temporal different resolution remotely sensed images

The mixed pixel problem affects the extraction of land cover information from remotely sensed images. Super-resolution mapping (SRM) can produce land cover maps with a finer spatial resolution than the remotely sensed images, and reduce the mixed pixel problem to some extent. Traditional SRMs solely adopt a single coarse-resolution image as input. Uncertainty always exists in resultant fine-resolution land cover maps, due to the lack of information about detailed land cover spatial patterns. The development of remote sensing technology has enabled the storage of a great amount of fine spatial resolution remotely sensed images. These data can provide fine-resolution land cover spatial information and are promising in reducing the SRM uncertainty. This paper presents a spatial–temporal Hopfield neural network (STHNN) based SRM, by employing both a current coarse-resolution image and a previous fine-resolution land cover map as input. STHNN considers the spatial information, as well as the temporal information of sub-pixel pairs by distinguishing the unchanged, decreased and increased land cover fractions in each coarse-resolution pixel, and uses different rules in labeling these sub-pixels. The proposed STHNN method was tested using synthetic images with different class fraction errors and real Landsat images, by comparing with pixel-based classification method and several popular SRM methods including pixel-swapping algorithm, Hopfield neural network based method and sub-pixel land cover change mapping method. Results show that STHNN outperforms pixel-based classification method, pixel-swapping algorithm and Hopfield neural network based model in most cases. The weight parameters of different STHNN spatial constraints, temporal constraints and fraction constraint have important functions in the STHNN performance. The heterogeneity degree of the previous map and the fraction images errors affect the STHNN accuracy, and can be served as guidances of selecting the optimal STHNN weight parameters.     

Cross-sensor calibration between Ikonos and Landsat ETM+ for spectral mixture analysis

Spectral mixture analysis is an algorithm that is developed to overcome the weakness in traditional land-use/land-cover (LULC) classification where each picture element (pixel) from remote sensing is assigned to one and only one LULC type. In reality, a remotely sensed signal from a pixel is often a spectral mixture from several LULC types. Spectral mixture analysis can derive subpixel proportions for the endmembers from remotely sensed data. However, one frequently faces the problem in determining the spectral signatures for the endmembers. This study provides a cross-sensor calibration algorithm that enables us to obtain the endmember signatures from an Ikonos multispectral image for spectral mixture analysis using Landsat ETM+ images. The calibration algorithm first converts the raw digital numbers from both sensors into at-satellite reflectance. Then, the Ikonos at-satellite reflectance image is degraded to match the spatial resolution of the Landsat ETM+ image. The histograms at the same spatial resolution from the two images are matched, and the signatures from the pure pixels in the Ikonos image are used as the endmember signatures. Validation of the spectral mixture analysis indicates that the simple algorithm works effectively. The algorithm is not limited to Ikonos and Landsat sensors. It is, in general, applicable to spectral mixture analysis where a high spatial resolution sensor and a low spatial resolution sensor with similar spectral resolutions are available as long as images collected by the two sensors are close in time over the same place.     

Change in habitat selection by Japanese macaques (Macaca fuscata) and habitat fragmentation analysis using temporal remotely sensed data in Niigata Prefecture,Japan

The aim of this study was to evaluate changes in macaque habitat selection during a 29-year period. We focused on the 1970s, when little crop damage was caused by Japanese macaques (Macaca fuscata), and the 2000s, when the damage became remarkable. Landsat/MSS from 1978 and ALOS/AVNIR-2 from 2007 were employed for land-cover mapping. For the 2007 land-cover classification, we applied an object-oriented image classification and a classification and regression tree. The Kappa coefficient of the 2007 land-cover map was 0.89. For the 1978 land-cover classification, change detection using principal component analysis and object-oriented image classification were applied to reduce resolution difference errors. The Kappa coefficient of the 1978 land-cover map was 0.84. We applied a Random Forest model for machine learning and data mining to predict the habitat selection of macaques. Several important environmental factors were identified for macaque habitat selection: the ratio of coniferous forest to farmland, distance to farmland, and maximum snow depth. The Random Forest model was extrapolated to the 1978 land-cover map. Over the 29-year period, coniferous forest changed to broad-leaved forest and/or mixed forest within the macaque habitat area. Coniferous forests were not selected as food resources by Japanese macaques. Furthermore, large-scale patches of farmland were used as food resources over the 29-year period. These changes indicated that habitat selection by Japanese macaques changed over the study period. The results show that the home range of macaques expanded, and macaques may now be distributed over a wider area as a result of changes in landscape configuration. Thus, forest planning, such as sustainable management of artificial conifer forests, is important for reducing crop damage.     

多尺度分割的高分辨率遥感影像变化检测

针对高空间分辨率的遥感影像,提出了一种基于多尺度分割的变化检测算法。采用Mean-Shift分割算法对影像进行多尺度分割,构建了不同尺度上的地理对象,以不同尺度上的地理对象灰度均值构建了变化检测的多尺度特征向量,采用变化矢量分析法获得最后的变化检测结果。以城镇区和农田区的Quick Bird影像对本文算法进行了检验,从精度评价的效果来看,无论城镇区还是农田区,采用面向对象的变化检测方法精度都高于基于单像素的检测方法,且当尺度层数固定时,多尺度组合的变化检测结果优于单一尺度的变化检测结果,对城镇、农田区域的变化检测的精度分别达到87.57%和81.55%。本文算法既可以顾及大面积同质区域变化,又可以反映小的地物目标及边缘部分的变化,能够很好地满足城镇、农田等不同环境背景下的变化检测需求,在国土资源监测中具有一定的应用价值。     

Integration of full-waveform LiDAR and hyperspectral data to enhance tea and areca classification

Hyperspectral image and full-waveform light detection and ranging (LiDAR) data provide useful spectral and geometric information for classifying land cover. Hyperspectral images contain a large number of bands, thus providing land-cover discrimination. Waveform LiDAR systems record the entire time-varying intensity of a return signal and supply detailed information on geometric distribution of land cover. This study developed an efficient multi-sensor data fusion approach that integrates hyperspectral data and full-waveform LiDAR information on the basis of minimum noise fraction and principal component analysis. Then, support vector machine was used to classify land cover in mountainous areas. Results showed that using multi-sensor fused data achieved better accuracy than using a hyperspectral image alone, with overall accuracy increasing from 83% to 91% using population error matrices, for the test site. The classification accuracies of forest and tea farms exhibited significant improvement when fused data were used. For example, classification results were more complete and compact in tea farms based on fused data. Fused data considered spectral and geometric land-cover information, and increased the discriminability of vegetation classes that provided similar spectral signatures.     

结合空间上下文与慢特征分析的多光谱影像变化检测

为了提高多光谱影像变化检测的精度,本文提出了一种结合空间上下文与慢特征分析的方法。首先采用自适应空间上下文提取算法围绕像素构建自适应区域,探索像素周围的上下文信息;然后通过迭代慢特征分析,由相应像素周围的成对自适应区域定量计算成对像素之间的变化强度,增强变化区域与未变区域的可分性;最后生成变化强度图像,采用大津阈值法作二值分类,将变化强度图划分为二值变化检测图。利用Landsat 7卫星ETM+传感器的图像,与4种基于代数的方法及基于变换的方法进行对比试验,结果表明,本文方法在降低漏检方面有所改善,提高了召回率。     

高分辨率遥感影像土地利用变化检测方法研究

提出一种利用高分辨率遥感影像进行土地利用变化检测的方法。以土地利用图为辅助数据,通过土地利用图和遥感影像的配准套合,获取影像像斑;同时,对遥感影像进行基于像素的监督分类,获取概略的类别图;再根据像斑内像素的类别编码完成子像斑的划分。以子像斑为影像分析的基本单位提取特征,以相关系数为相似性测度衡量不同时期子像斑的特征相似性,用ROC曲线(接受者操作特性曲线)代替经验选取的方法自动获取变化阈值,确定像斑是否发生变化。以武汉市区局部QuickBird 2002年和2005年多光谱影像、相同地区2002年1∶10 000土地利用图为实验数据进行了算法的实验,结果显示绝大部分的变化区域都可以被提取出来,实验方法可行。     

The ‘Stained Glass Procedure’, a new method to compare classification performance of images acquired with different pixel sizes

Objective comparison of classification performance of earth observation images, acquired at different spatial resolutions (e.g. NOAA-AVHRR, IRS-MOS, IRS-WiFS, Landsat-TM, IRS-LISS), is complicated because both class definition and training site selection are hampered by the inherent scale differences. This paper presents a new, generic method to compare the information content of such a set of images, the “Stained Glass Procedure”. It overcomes the stated problems by computing the scale-dependent, internal spectral variation in an image and by using this as an indicator for land cover information. The Stained Glass Procedure creates segments in the images and calculates the internal spectral variation in a high-spatial-resolution image for each segment. For each image from the set the average variance, weighted to area, is calculated. The Stained Glass Procedure can be used to predict the performance of sensors that are not available, yet, or to roughly determine the optimal spatial resolution for the classification of a specific area.The procedure was applied to images with pixel sizes ranging from 23 to 1100 m. Classification detail of Envisat-MERIS (300 m pixel size), not included in the image set, could be predicted accurately using the Stained Glass Procedure.The Stained Glass Procedure applies one procedure to all images, without any subjective decision during the analysis, thus offering a method to compare images with different pixel sizes in terms of classification detail that is truly objective.     

Landsat TM数据不同辐射校正方法对土地覆盖遥感分类的影响

本文主要是探索Landsat TM数据不同辐射校正方法对土地覆盖遥感分类的影响。介绍了使用的3种不同辐射校正方法(ATCOR3、FLAASH以及查找表)和两种分类算法。在分类实验部分,根据样本的地理坐标在3景校正影像中分别采集训练样本并训练各自的分类器,并交叉用于其他辐射校正影像的土地覆盖遥感分类。实验结果表明:(1)用于分类器训练的样本采集自待分类影像时的分类精度明显高于采集自其他影像的分类精度;(2)3种辐射校正影像的分类结果存在差异,其中使用ATCOR3和FLAASH方法校正后影像的分类结果有更相近的精度;(3)辐射校正对分类类别的影响不同,其中对森林类型影响最大,对裸地等其他类别影响相对较小。     

基于最小距离法的面向对象遥感影像分类

主要介绍了集成基于对象的影像分析与最小距离分类方法的原理,采用中卫市World ViewⅡ影像进行土地覆盖分类研究,并将分类结果与传统的基于像元的最小距离分类结果进行对比。目视解译与定量评价均表明:基于对象方法的各项指标更优越,总体精度由0.85提高到0.87,Kappa系数由0.81提高到0.84。因此,对于高分辨率遥感影像,集成最小距离分类器,基于对象的信息提取方法要优于基于像元方法,分类结果精度更高。     

Evaluation of Goddard’s LiDAR,hyperspectral, and thermal data products for mapping urban land-cover types

Goddard’s LiDAR (Light Detection And Ranging), hyperspectral and thermal (G-LiHT) airborne imager is a new system to advance concepts of data fusion for worldwide applications. A recent G-LiHT mission conducted in June 2016 over an urban area opens a new opportunity to assess the G-LiHT products for urban land-cover mapping. In this study, the G-LiHT hyperspectral and LiDAR-canopy height model (LiDAR-CHM) products were evaluated to map five broad land-cover types. A feature/decision-level fusion strategy was developed to integrate two products. Contemporary data processing techniques were applied, including object-based image analysis, machine-learning algorithms, and ensemble analysis. Evaluation focused on the capability of G-LiHT hyperspectral products compared with multispectral data with similar spatial resolution, the contribution of LiDAR-CHM, and the potential of ensemble analysis in land-cover mapping. The results showed that there was no significant difference between the application of the G-LiHT hyperspectral product and simulated Quickbird data in the classification. A synthesis of G-LiHT hyperspectral and LiDAR-CHM products achieved the best result with an overall accuracy of 96.3% and a Kappa value of 0.95 when ensemble analysis was applied. Ensemble analysis of the three classifiers not only increased the classification accuracy but also generated an uncertainty map to show regions with a robust classification as well as areas where classification errors were most likely to occur. Ensemble analysis is a promising tool for land-cover classification.     

结合L0平滑和超像素的高空间遥感影像非监督分类

针对基于像元的非监督分类方法对高空间遥感影像分类时易形成“椒盐”噪声和产生大量错分、漏分的问题,提出了一种结合L0平滑和超像素的非监督分类方法．首先采用L0算法对高空间遥感影像进行平滑操作,减少大量图像噪声及冗余信息;然后采用简单的线性迭代聚类（SLIC）超像素方法处理平滑后图像,进一步抑制椒盐现象的同时降低处理复杂度,得到初始聚类图;最后采用K-means非监督分类方法得到最终分类结果图．为验证本文提出的方法,选取3景高空间遥感影像作为实验数据．试验结果表明,采用提出的方法能准确对地物分类,且总体精度分别达到了72.46％、77.55％和78.44％,Kappa系数分别达到0.788、0.779和0.779．提出方法能有效解决分类中存在的“椒盐”现象,可提高分类精度,对高空间遥感影像分类具有一定的参考价值．      

Development of a remotely sensed,historical land-cover change database for rural Chihuahua,Mexico

The purpose of this article is to describe the development of a remotely sensed, historical land-cover change database for the northwestern quarter of Chihuahua, Mexico, The database consists of multi-temporal land-cover classifications and change detection images. The database is developed to facilitate future investigations that examine urban–rural linkages as possible drivers of rural land-use and land-cover changes. To develop the needed land-cover change database, this study uses the North American Landsat Characterization (NALC) MSS triplicates because of their temporal depth and spatial breadth. Challenges exist, however, to effective classification and change detection using the NALC triplicates, including illumination differences across multiple scenes and periods caused by topographic and solar variations and the lack of ground reference data for historic periods. Therefore, creation of the database is a four step process. First, extensive pre-processing is performed to enhance comparability of multi-date images. Pre-processing includes topographic correction, mosaic creation and multi-date radiance normalization. Second, ancillary sources of land-cover data are combined with visual interpretations of enhanced images to define reference pixels used to classify the images using the maximum likelihood algorithm. Third, classification accuracy is assessed. Fourth, post-classification change detection is performed. Results indicate significant image improvements after pre-processing that permit very good overall classification (86.26% classified correctly) and change detection. To conclude, findings are presented that indicate significant changes to arid grasslands/shrublands and forest resources in mountainous regions.     

面向对象的高光谱遥感影像稀疏表示分类

稀疏表示用于高光谱遥感影像分类多是基于像素层次来处理的。文中提出一种面向对象的高光谱遥感影像稀疏表示分类方法。首先从高光谱影像中提取4个波段组成标准的多波段影像,进行面向对象的影像分割;然后计算各对象在各波段上的光谱均值,并选取少量样本进行训练;最后利用基于Fisher字典学习的稀疏表示进行高光谱遥感影像的分类。实验结果表明,该方法可以利用较少的样本得到较好的分类效果,与基于像素层的稀疏分类相比较,分类精度与效率均有所提高,分类结果更接近真实地物,避免了零碎图斑。     

A Neural Network Landuse Classifier for SAR Images Using Textural and Fractal Information

Abstract

Artificial neural networks (ANN) have recently been popularly used in image classification. Input features to most ANNs are extracted based on a one class per pixel basis. This requires a large number of training samples and thus a slow training rate. In this paper, we describe the use of a windowing technique to extract textural features such as average intensity, second moment of intensity histogram and fractal surface dimension from an image. This method of image characterization reduces the number of training samples efficiently, yet retains a reasonable overall classification accuracy. The ANN is trained based on the back‐error propagation algorithm. The method is applied for landuse classification of Synthetic Aperture Radar (SAR) images. An example is given for a site in Kedah State, Malaysia. The SAR images (HH,HV,VV) were taken by the Canadian Centre for Remote Sensing (CCRS) CV‐580 airborne C‐band SAR system in November 1993 during their GlobeSAR mission in Malaysia. These multi‐polarization SAR images are co‐registered with a Landsat Thematic Mapper (TM) channel 5 image from same area. An overall classification accuracy of about 86.95% is achieved using windowing technique, as compared to 68.22% based on one class per pixel approach. This shows that through fractal and textural information, the windowing technique when applied in an ANN classifier has a great potential in remote sensing applications.     

样本迁移支持下的遥感影像自动分类方法

迁移学习是运用已有知识对相关的不同领域的问题进行求解的一种机器学习方法,本文结合这一方法,提出了一种基于先验知识的样本自动选取方法,并构建了一套土地覆盖自动分类的算法框架。该方法主要面向Landsat数据,通过图像变化检测技术与光谱形状编码的方法,从源领域中迁移适用的地物类别知识并标记在目标影像中,使用SVM完成基于样本迁移的自动分类流程。结果表明,该方法可以获得可靠的自动分类结果,一定程度上满足遥感信息的大范围提取与长时间序列处理分析的发展需求。     

面向对象的高分辨率遥感影像土地覆盖信息提取

利用高分辨率影象提取土地覆盖信息的关键技术在于如何利用丰富的纹理信息来弥补光谱信息的不足。面向对象的图像分类技术改变了传统的面向像素的分类技术:(1)用来解译图像的信息并不在单个像元中,而是在图像对象和其相互关系中;采用多分辨率对象分割方法生成图像对象,提高了分类信息的信噪比;基于对象的分类技术不同于纯粹的光谱信息分类,图像对象还包含了许多的可用于分类的一些其他特征:形状、纹理、相互关系、上下关系等信息。面向对象的土地覆盖分类结果与传统分类方法相比,其特征提取算子更加地适合于几何信息和结构信息丰富的高分辨率图像的自动识别分类。