Optimizing unsupervised classifications of remotely sensed imagery with a data-assisted labeling approach

The quality of remotely sensed land use and land cover (LULC) maps is affected by the accuracy of image data classifications. Various efforts have been made in advancing supervised or unsupervised classification methods to increase the repeatability and accuracy of LULC mapping. This study incorporates a data-assisted labeling approach (DALA) into the unsupervised classification of remotely sensed imagery. The DALA-unsupervised classification algorithm consists of three steps: (1) creation of N spectral-class maps using Iterative Self-Organizing Data Analysis Technique Algorithm (ISODATA); (2) development of LULC maps with assistance of reference data; and (3) accuracy assessments of all the LULC maps using independent reference data and selection of one LULC map with the highest accuracy. Classification experiments with a composite image of a Landsat Thematic Mapper (TM) image and an Enhanced Thematic Mapper Plus (ETM+) image suggest that DALA was effective in making unsupervised classification process more objective, automatic, and accurate. A comparison between the DALA-unsupervised classifications and some conventional classifications suggests that the DALA-unsupervised classification algorithm yielded better classification accuracies compared to these conventional approaches. Such a simple, effective approach has not been systematically examined before but has great potential for many applications in the geosciences.     

Aboveground biomass estimation in intervened and non-intervened Nothofagus pumilio forests using remotely sensed data

Forest inventory data can be used along with remotely sensed data to estimate biomass and carbon stocks over large and inaccessible forested areas. In this study, the relationship between satellite-derived multispectral data and forest variables from intervened and non-intervened Nothofagus pumilio forest stands located in the Magellan region of Chile was examined, in order to quantify the over bark volume (OBV) and aboveground tree biomass (AGTB). Four vegetation parameters – the green normalised difference vegetation index (GNDVI), normalised difference vegetation index (NDVI), simple ratio (SR) and vegetation cover fraction (VCF) – were retrieved from an Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) image of the study area. The results indicate that only the VCF presents significant differences among intervened and non-intervened stands. The best OBV and AGTB models (R² = 0.58) were found using the SR index and the VCF as predictors. This result could be transferred to estimate biomass and volume in other Nothofagus pumilio forests with similar conditions. Moreover, it can be used to assess temporal carbon changes.     

适合高分辨率遥感影像处理的分布式环境研究

随着遥感卫星的发展及遥感图像分辨率的增加,遥感图像的分布式处理变得越来越重要,应用分布式环境处理高分辨率遥感图像变得越来越迫切。传统的分布式处理主要采用B／S或C／S结构,这种结构当多用户并发访问时往往由于服务器的访问量过大而容易造成网络堵塞。在本项研究中采用了当前比较先进的网格计算及Web Service中资源共享的思想,根据高分辨率遥感图像分布式处理方面的需要,对遥感图像分布式处理框架进行了设计,并结合实际应用中的遥感图像分割对系统进行了实现。结果表明,新模型支持下的系统在多用户并发访问支持及系统优化方面都有很大的改善。     

遥感影像亚像元定位研究综述

遥感影像亚像元定位是在混合像元分解基础上,利用地物空间分布特征确定不同地物类型在混合像元中的具体位置,得到亚像元尺度的地物分类图,是一种有效解决混合像元空间不确定性的方法。首先介绍遥感影像亚像元定位的基本概念,分析亚像元定位的理论模型和求解算法;然后总结亚像元定位模型的误差来源、精度评价方法以及结果不确定性的表达手段,同时讨论利用辅助数据源提高亚像元定位精度的主要方法;最后对亚像元定位的研究趋势做了进一步展望。     

颜色保真的遥感图像融合方法的研究

针对遥感图像融合Brovey变换法存在颜色失真的现象,提出了一种低通比值融合法,该融合法首先对高几何分辨率的全色波段进行低能滤波,尔后将低分辨率多光谱图像与全色波段图像相乘,再除以滤波后的全色波段图像,便得到融合图像。从辐照的角度证明了该低通比值融合法具备理论基础,并从目视评价、定量分析、分类精度证实了该低通比值融合法优于Brovey变换法,该低通比值融合法是一种能较好地保全低分辨率多光谱图像颜色的融合方法。     

Image processing for boundary extraction of remotely sensed data

Image processing algorithms for analysing remotely sensed data are developed. The algorithms proposed in the paper provide means for autoregressive texture modelling and for boundary detection of uniform subimage areas. The boundary detection methods make use of a semicircular entropy operator and of the binary hypothesis testing of the Poisson counting process. The proposed algorithms are applied to the pattern analysis of the isothermal distribution in the oceanic environment.     

基于遥感图像的颜色保真融合方法的研究*

针对遥感图像融合Brovey变换法存在颜色失真的现象,提出了一种低通比值融合法。该融合法首先对高几何分辨率的全色波段进行低通滤波,然后将低分辨率多光谱图像与全色波段图像相乘,再除以滤波后的全色波段图像,便得到融合图像。从辐照的角度证明了该低通比值融合法具备理论基础,并从目视评价、定量分析、分类精度证实了该低通比值融合法优于Brovey变换法,该低通比值融合法是一种能较好地保全低分辨率多光谱图像颜色的融合方法。     

Topographic correction algorithm for remotely sensed data accounting for indirect irradiance

The solar irradiance incidents upon terrain surface are composed of three parts, i.e. direct solar irradiance, diffuse sky irradiance and reflected irradiance from the adjacent surface, respectively. Most of the topographic correction models only account for the topographic effect induced from direct solar irradiance, and few models take the topographic effects from the last two components of solar irradiance into account. A physically based topographic correction algorithm aiming to overcome this shortcoming, called a three-factor correction model, was developed based on theoretical analysis of radiation transferring processes along an undulating surface, atmosphere and satellite sensor geometry under the assumption of Lambertian surface. On the basis of this three-factor correction model, an advanced algorithm accounting for the bidirectional reflectance distribution function (BRDF) nature of non-Lambertian surface, called the three-factor+C topographic correction model, was developed by introducing an empirical parameter C to approximate the indirect irradiance contribution of non-Lambertian surface. Performances of these two newly developed algorithms were tested and compared with those of Cosine and C correction algorithms for a selected rugged terrain on the south flank of the Qinling Mountain, China. Visual comparison and statistical analysis were adopted for quantitative evaluation on topographic corrections of a Landsat-7 Enhanced Thematic Mapper Plus (ETM+) image in the study. The results suggested that the general performance of the algorithms for topographic correction ranks the three-factor+C correction, C correction, three-factor correction and Cosine correction from excellent to poor in order, which implies the promising potential of the proposed algorithms in effective topographic correction applications in remote sensing techniques.     

Symbolic agglomerative clustering for quantitative analysis of remotely sensed data

An efficient nonparametric, hierarchical, symbolic agglomerative clustering procedure based on the mutual nearest neighbourhood concept is proposed for classifying remotely sensed multispectral data. The procedure utilized a data reduction technique and an innovative symbolic concept to minimize the memory and computational time requirements. A new non-metric similarity measure and a novel method of formulation of composite symbolic objects are proposed to enrich the performance of the algorithm. A Mean Difference Index (MDI) concept for identifying the optimal number of classes was used. Experiments were conducted on IRS (Indian Remote Sensing) satellite data to authenticate the efficacy of the procedure.     

Statistical evaluation of remotely sensed data for water quality monitoring

The primary objective of this study was to determine relationships between water quality parameters (WQPs) and digital data from the Landsat satellite to estimate and map the WQP in the Porsuk Dam reservoir. Suspended sediments (SS), chlorophyll a (chl-a), NO₃-N and transmitted light intensity depth (TLID) were the parameters for water quality determination used in this study. Collection of these data, obtained from the General Directorate of State Hydraulic Works (GDSHW) was synchronized with the Landsat satellite overpass of the September 1987. The relationships between the brightness values (BV) of the TM data and WQP were determined. Using the TM data, we developed multiple regression equations to estimate the WQPs, and the validation of these equations was checked by using ANOVA. The effects of SS, NO₃-N and chl-a on TLID were tested not only for ground data, but also for TM datasets. Regression equations were developed for two different datasets and the homogeneity of those equations was tested. Finally, these regression equations evaluated from digital TM data and ground data were applied to map TLID values.     

遥感影像超分辨率制图研究进展

尺度问题是土地覆盖分类中的一个核心问题,向下尺度转换又是其中的难点。混合像元分解可以得到亚像元尺度的类别组分百分比,但无法求得亚像元的具体位置。遥感影像超分辨率制图是由粗空间分辨率的影像得到高空间分辨率分类结果图的技术,可用于地表分类向下尺度转换,近年来该技术已成为遥感影像分类和尺度转换领域的研究热点。对超分辨率制图研究进展做了详细论述,从超分辨率制图的发展和研究现状、主要方法、精度评价等几方面进行了详细阐述,并分析了当前超分辨率制图算法存在的主要问题,以及可能的研究重点和发展空间。     

Comparison and improvement of methods for identifying waterbodies in remotely sensed imagery

This article first examines three existing methods of delineating open water features, i.e. the normalized difference water index (NDWI), the modified normalized difference water index (MNDWI) and a method combining the near-infrared (NIR) band and the maximum likelihood classification. We then propose two new methods for the fast extraction of water features in remotely sensed imagery. Our first method is a pixel-based procedure that utilizes indices and band values. Based on their characteristic spectral reflectance curves, waterbodies are grouped into three types – clear, green and turbid. We found that the MNDWI is best suited for identifying clear water. Green water has its maximum reflectance in Landsat Thematic Mapper (TM) band 4 (NIR band), whereas turbid water has its maximum reflectance in TM band 5 (mid-infrared band). Our second method integrates our pixel-based classification with object-based image segmentation. Two Landsat scenes in Shaanxi Province, China, were used as the primary data source. Digital elevation models (DEMs) and their derived slope maps were used as ancillary information. To evaluate the performance of the proposed methods, extraction results of the three existing methods and our two new methods were compared and assessed. A manual interpretation was made and used as reference data. Results suggest that our methods, which consider the diversity of waterbodies, achieved better accuracy. Our pixel-based method achieved a producer's accuracy of 92%, user's accuracy of 90% and kappa statistics of 0.91. Our integrated method produced a higher producer's accuracy (95%), but a lower user's accuracy (72%) and kappa statistics (0.72), compared with the pixel-based method. The advantages and limitations of the proposed methods are discussed.     

A new algorithm for remotely sensed image texture classification and segmentation

In this paper, we propose a new algorithm for remotely sensed image texture classification and segmentation. We observe that the traditional method least square error (LSE) is unstable in practical applications. This motivates us to develop a more stable method. We have proposed the regularization technique to suppress the instability of LSE in previous research. Our contribution in this paper is that we propose a new stable method, which is based on the total variation (TV) for reducing instability in texture analysis, and apply it to remotely sensed image texture classification and segmentation. Experimental results on remotely sensed images demonstrate that our new algorithm is superior to LSE and seems promising in applications.     

Unmixing-based content retrieval system for remotely sensed hyperspectral imagery on GPUs

This paper presents a new unmixing-based retrieval system for remotely sensed hyperspectral imagery. The need for this kind of system is justified by the exponential growth in the volume and number of remotely sensed data sets from the surface of the Earth. This is particularly the case for hyperspectral images, which comprise hundreds of spectral bands at different (almost contiguous) wavelength channels. To deal with the high computational cost of extracting the spectral information needed to catalog new hyperspectral images in our system, we resort to efficient implementations of spectral unmixing algorithms on commodity graphics processing units (GPUs). Spectral unmixing is a very popular approach for interpreting hyperspectral data with sub-pixel precision. This paper particularly focuses on the design of the proposed framework as a web service, as well as on the efficient implementation of the system on GPUs. In addition, we present a comparison of spectral unmixing algorithms available in the system on both CPU and GPU architectures.     

Stepwise-then-intelligent algorithm (STIA) for optimizing remotely sensed image rectification

ABSTRACT

To address the problems of parameter selection and accuracy optimization of models in image rectification, this article first proposes a novel stepwise-then-intelligent algorithm (STIA) for image rectification optimization, which includes the following steps. First, stepwise regression is suggested to simultaneously solve the over-parameterization problem and select the optimum parameters of the polynomial model and rational function model according to different terrains. Second, intelligent algorithms, e.g. the genetic algorithm (GA) and particle swarm optimization (PSO), are proposed to search for better results based on an innovative search range determined by the uncertainty propagation and 3-sigma rule. The experimental results show that the proposed STIA can achieve higher accuracy than conventional methods; and in most cases, the PSO algorithm used in STIA is superior to the GA used in STIA in measures of time and accuracy. Moreover, stepwise-then-PSO algorithm exhibits the best performance of all compared methods, including least squares, stepwise regression, total least squares and partial least squares.     

A modified GAC model for extracting waterline from remotely sensed imagery

At present many methods are available for extracting waterline automatically from remotely sensed imagery. They are commonly limited by their inability to accurately detect waterline when the land–sea interface is not so distinctive on the imagery. This study aims to develop a new model of extraction in which both the spectral features of water and the spatial feature of waterline itself are incorporated into the level set framework. This model consists of three terms. The first term, which is derived from the spectral features, drives the curve evolution. The second term, curvature flow, guarantees the smoothness of the curve in the process of evolution. The third term, the boundary attraction force, attracts the evolving curve to the genuine position. Experimental results with Landsat Thematic Mapper (TM) data sets validate the effectiveness of the proposed model. It can improve the accuracy of extracted waterline and effectively prevent the omission of weak boundaries from the detected results that are common in both geodesic active contour (GAC) and distance-regularized level set evolution (DRLSE) results.     

A fast separability-based feature-selection method for high-dimensional remotely sensed image classification

Because of the difficulty of obtaining an analytic expression for Bayes error, a wide variety of separability measures has been proposed for feature selection. In this paper, we show that there is a general framework based on the criterion of mutual information (MI) that can provide a realistic solution to the problem of feature selection for high-dimensional data. We give a theoretical argument showing that the MI of multi-dimensional data can be broken down into several one-dimensional components, which makes numerical evaluation much easier and more accurate. It also reveals that selection based on the simple criterion of only retaining features with high associated MI values may be problematic when the features are highly correlated. Although there is a direct way of selecting features by jointly maximising MI, this suffers from combinatorial explosion. Hence, we propose a fast feature-selection scheme based on a ‘greedy’ optimisation strategy. To confirm the effectiveness of this scheme, simulations are carried out on 16 land-cover classes using the 92AV3C data set collected from the 220-dimensional AVIRIS hyperspectral sensor. We replicate our earlier positive results (which used an essentially heuristic method for MI-based band-selection) but with much reduced computational cost and a much sounder theoretical basis.     

Directional morphological image transforms for lineament extraction from remotely sensed images

Morphological image transforms find their basis in the notions of mathematical morphology. An attempt is made to develop morphological transforms for enhancement of directional edges and lineaments from satellite imagery. Geostructural features are generally oriented in preferred directions. This information has been utilized in designing structuring elements for morphological transforms. Several shapes and sizes of structuring elements have been designed and applied to delineate lineaments in different litho-environments of the same study area. Top hat transform has been applied using the directional structuring elements for edge image at 0^o, 30^o, 60^o, 90^o, 120^o and 150^o. Geological lineaments such as faults could be easily identified on an edge image obtained using top hat transformation followed by an image superimposition technique. The lineament map was developed using directionally enhanced edge images and the results have been verified.     

Genetic classifiers for remotely sensed images: Comparison with standard methods

In this article the effectiveness of some recently developed genetic algorithm-based pattern classifiers was investigated in the domain of satellite imagery which usually have complex and overlapping class boundaries. Landsat data, SPOT image and IRS image are considered as input. The superiority of these classifiers over k-NN rule, Bayes' maximum likelihood classifier and multilayer perceptron (MLP) for partitioning different landcover types is established. Results based on producer's accuracy (percentage recognition score), user's accuracy and kappa values are provided. Incorporation of the concept of variable length chromosomes and chromosome discrimination led to superior performance in terms of automatic evolution of the number of hyperplanes for modelling the class boundaries, and the convergence time. This non-parametric classifier requires very little a priori information, unlike k-NN rule and MLP (where the performance depends heavily on the value of k and the architecture, respectively), and Bayes' maximum likelihood classifier (where assumptions regarding the class distribution functions need to be made).     

Displaying remotely sensed vegetation dynamics along natural gradients for ecological studies

Normalized difference vegetation index (NDVI) datasets are growing in popularity to represent vegetation dynamics in ecological studies. Because of its multidimensional nature, it is difficult to visualise the spatial and temporal components of NDVI datasets simultaneously. This letter presents a method to display vegetation dynamics as captured by the NDVI along natural gradients and to visualise and test correlations between vegetation phenology and animal movement.