ENDMEMBER EXTRACTION FOR HYPERSPECTRAL IMAGES USING WATERSHED AND NORMALIZED CUTS

Several methods based on satellite data products are available to estimate snow cover properties, each one with its pros and cons. This work proposes and implements a novel methodology that integrates three main processes applied to MODIS satellite data for snow cover property reconstruction: (1) the increase in the spatial resolution of MODIS (MOD09) data to 250 m using a spectral fusion technique; (2) a new proposal of snow-cloud discrimination; (3) the daily spatio-temporal reconstruction of snow extent and its albedo signature using the endmembers extraction and spectral mixture analyses. The snow cover reconstruction method was applied to the Brunswick Peninsula, Chilean Patagonia, a low-elevation (<1500 m a.s.l.) mid-latitude area. The results show a 98% agreement between MODIS snow detection and ground-based snow measurements at the automatic weather station, Tres Morros (53.3174°S, 71.2790°W), with fractional snow cover values between 20% and 50%, showing a close relationship between snow and vegetation type. The number of snow days compiled from the MODIS data indicates a good performance (Pearson’s correlation of 0.9) compared with the number of skiing days at the Cerro Mirador ski center, Punta Arenas. Although the number of seasonal snow days showed a significant increasing trend of 0.54 days/year in the Brunswick Peninsula during the 2000–2020 period, a significant decrease of −4.64 days/year was detected in 2010–2020.

Most spectral mixture analyses in the literature overlook the spatial correlation of neighborhood pixels. The main contribution of this paper is to consider the impacts of both spatial and spectral information prior to endmember (EM) extraction algorithms. Hence, we take advantage of a top-down over-segmentation algorithm in combination with fuzzy c-means (FCM) clustering to identify spatially homogenous over-segments with minimum spectral variability and high spatial correlation. FCM provides a soft segmentation while its partial membership matrix is exploited to calculate a novel local entropy criterion (LEC) at pixels seated in homogenous over-segments. Afterwards, by performing an adaptive threshold per homogenous over-segment, pixels with high LEC values which have high certainty to associate with only one class are selected as pure ones. LEC calculations lead to preserving level of unmixing accuracy while speeding up EM extraction. This subject is important for large images particularly with real-time limitations. With respect to experiments accomplished on synthetic and AVIRIS hyperspectral images, clustering, over-segmentation, and entropy preprocessing has a simple and fast framework while it relatively outperforms the state-of-the-art procedures in terms of extraction accuracy and computing time.

Abstract. An essential step in the characterization of surface materials using hyperspectral image analysis is image classification using endmembers. Spectral unmixing is the best method for hyperspectral image classification. This method assumes that the pixel-topixel variability in scene results from varying proportions of spectral endmembers. Spectral endmembers can be derived from the imagery or measurements in the laboratory or field. The primary objective of this paper was to assess the ability of extracted endmembers against some different solutions for extraction of endmembers for hyperspectral image classification. In this paper, we compared the Pixel Purity Index (PPI) and the Sequential Maximum Angle Convex Cone (SMACC) as two popular methods of endmember extraction with library and field spectral. We used spectral information divergence for detection desirable endmembers from field spectral. For accuracy assessment of spectral mixture analysis and production of endmember abundance images for each of methods, the linear spectral unmixing algorithm is used. After a comparison between the results of these methods, it has been verified that field spectral have a better classification result in comparing with other endmember extraction methods. Also, the PPI has reliable results as an automatic endmember extraction method in comparing.

Endmember extraction (EE) is defined as the process of identifying a set of pure and unique spectral signatures of the materials combined at pixel level in a remotely sensed hyperspectral scene. These extracted signatures are utilized to decompose scene into abundance fractions by means of Spectral Mixture Analysis (SMA) model Most of algorithms used for EE, are established on spectral discrimination without any attention to spatial context and distribution of image pixels. Some years ago several algorithms are implied which applies spatial context in addition to spectral discrimination in a distinct module such as SPP and SSPP. In this paper, we propose a novel Cluster-based Spatial Border Removal Preprocessor (CSBRP) by removing spatial borders of cluster map and calculate spectral purity weight of residual pixels in order to look for spatially homogenous and spectrally pure pixels with help of spectral purity weight threshold in a distinct module. The purpose of CSBRP is to improve RMSE reconstruction of original image and a new metric defined as Efficiency by reducing the processing time needed for EE. Experimental results on the proposed preprocessor outperform the state of the art EE methods, SPP and SSPP in real Aviris hyperspectral Salinas and Indian Pines images.

Endmember extraction is an important step in spectral mixture analysis when endmembers are unknown. Endmembers are usually assumed to be pure pixels present in an image scene. Under this circumstance, endmember extraction is to find the most distinctive pixels. To make the searching process more efficient, the sequential forward search (SFS) method is generally used, where the next endmember is determined with a certain criterion based on the currently extracted endmember set. This letter proposes a new criterion which is related to the estimated endmember abundances. Compared to other sequential endmember extraction algorithms, the proposed method can find all the different endmembers faster. This letter also proposes to use the sequential forward floating search method as the substitute of SFS, which can improve the performance of all the sequential endmember extraction algorithms.

Random Projection (RP) has been proven to be a powerful technique for Dimension Reduction (DR). In this paper, it is applied to hyperspectral images as a DR preprocess step for Endmember Extraction (EE). Theoretical analysis demonstrates that RP can preserve geometric simplex fitting by hyperspectral data perfectly. Therefore, endmembers, which play an extremely important role for Spectral Mixture Analysis (SMA) of hyperspectral images, can be extracted from the projected data in a subspace by RP and the computational complexity of EE can be greatly reduced. Experimental results demonstrate that RP is computational efficient and data-independent DR technique for EE.

Endmember extraction is an important task for hyperspectral analysis; the accurate identification of endmembers enables efficient spectral unmixing and classification. In the paper, a new endmember extraction algorithm based on a modified MSS approach with LP error as initial spectrum selection algorithm, and OPD measure as similarity is proposed. The endmembers extracted by modified MSS are more similar than that of MSS algorithm; from the experiments results, it has proved that our proposed method outperforms the existed MSS and N-FINDR algorithms.

This study assessed the feasibility of spectral mixture analysis (SMA) of Landsat thematic mapper (TM) data for monitoring estuarine vegetation at species level. SMA modelling was evaluated, using χ2 test, by comparing SMA fraction images with a precisely classified QuickBird image that has a higher spatial resolution. To clearly understand the strengths and weaknesses of SMA, eight SMA models with different endmember combinations were assessed. When the TM data dimension for SMA and the endmember number required were balanced, a model with three endmembers representing water and two vegetation types was most accurate, whereas a model with five endmembers approximated the actual surface situation and generated a relatively accurate result. Our results indicate that an SMA model with appropriate endmembers had relatively satisfactory accuracy in monitoring vegetation. However, errors might occur in SMA fraction images, especially in models with an inappropriate endmember combination, and the errors were mainly distributed in areas filled with water or near water. Therefore, short vegetation usually submerged during high tide tended to be poorly predicted by SMA models. These results strongly suggest that tide water has a great influence on SMA modelling, especially for short vegetation.

A method of incorporating the multi-mixture pixel model into hyperspectral endmember extraction is presented and discussed. A vast majority of hyperspectral endmember extraction methods rely on the linear mixture model to describe pixel spectra resulting from mixtures of endmembers. Methods exist to unmix hyperspectral pixels using nonlinear models, but rely on severely limiting assumptions or estimations of the nonlinearity. This paper will present a hyperspectral pixel endmember extraction method that utilizes the bidirectional reflectance distribution function to model microscopic mixtures. Using this model, along with the linear mixture model to incorporate macroscopic mixtures, this method is able to accurately unmix hyperspectral images composed of both macroscopic and microscopic mixtures. The mixtures are estimated directly from the hyperspectral data without the need for a priori knowledge of the mixture types. Results are presented using synthetic datasets, of multi-mixture pixels, to demonstrate the increased accuracy in unmixing using this new physics-based method over linear methods. In addition, results are presented using a well-known laboratory dataset.

Analysis of spectral mixtures is important in remote sensing imaging spectroscopy, because essentially the spectrum of any pixel of a natural scene is a mixture. The analysis of mixed spectra, known as Spectral Mixture Analysis (SMA) , is the subject of this chapter. SMA attempts to answer two questions: (a) What are the spectra of the individual materials? (b) What are the proportions of the individual materials? We focus on linear mixing because of its relative analytical and computational simplicity and because it works satisfactorily in many practical applications.We discuss the physical aspects of the linear mixing model, geometrical interpretations and algorithms, and statistical analysis using the theory of least squares estimation. The main applications of SMA are in the areas of hyperspectral image interpretation and subpixel target detection. Spectral Mixing When a ground resolution element contains several materials, all these materials contribute to the individual pixel spectrum measured by the sensor. The result is a composite or mixed spectrum, and the “pure” spectra that contribute to the mixture are called endmember spectra. Spectral mixtures can be macroscopic or intimate depending on what scale the mixing is taking place (see Figure 9.1). In a macroscopic mixture the materials in the field of view are optically separated in patches so there is no multiple scattering between components (each reflected photon interacts with only one surface material). Such mixtures are linear: that is, the combined spectrum is simply the sum of the fractional area times the spectrum of each component. Linear mixing is possible as long as the radiation from component patches remains separate until it reaches the sensor. In an intimate mixture, such as the microscopic mixture of mineral grains in a soil or rock, a single photon interacts with more than one material. In this case, mixing occurs when radiation from several surfaces combines before it reaches the sensor. These types of mixtures are nonlinear in nature and therefore more difficult to analyze and use. To illustrate some of the issues involved in SMA, we discuss some examples of mixed spectra provided by Adams and Gillespie (2006). Figure 9.2(a) shows spectra for mixtures of a material having a featureless spectrum (quartz) with a material having a spectrum with diagnostic absorption bands (alunite).

Spectral unmixing is an important problem for remotely sensed hyperspectral data exploitation. Automatic spectral unmixing can be viewed as a three-stage problem, where the first stage is subspace identification, the next one is endmember extraction, and the final one is abundance estimation. In this sequence, endmember extraction is the most challenging problem. Many researchers have attempted to extract endmembers from hyperspectral images using spectral information only. However, it is well known that the inclusion of spatial information can improve the endmember extraction task. In this article, we introduce a new endmember extraction algorithm that exploits both spectral and spatial information. A main innovation of the proposed algorithm is that spatial information is exploited using entropy, while spectral information is exploited using convex set optimization. In the literature, none of the spatial–spectral algorithms has used entropy as spatial information. The inclusion of this entropy-based spatial information improves the accuracy of the endmember extraction process. The results obtained by the proposed algorithm are compared (using a variety of metrics) with those obtained by other state-of-the-art methods, using both synthetic and real datasets. Our experimental results demonstrate that the proposed algorithm outperforms many available algorithms.

Impervious surfaces have been widely considered as the key indicator for evaluating urbanization and environmental quality. As one of the most widely applied methods, spectral mixture analysis (SMA) has been commonly used for mapping urban impervious surface fractions. When implementing SMA, the original multispectral remote-sensing reflectance images are served as the foundation and key to successful SMA. However, the limited spectral variances among different land covers from the original reflectance images make it challenging in information extraction and results in unsatisfactory mapping results. To address this issue, a new method has been proposed in this study to improve urban impervious surface mapping through integrating statistical methods and SMA. In particular, two traditional statistical methods, principal component analysis (PCA) and minimum noise fraction rotation (MNF) were applied to highlight the spectral variances among different land covers. Three endmember classes (impervious surface, soil, and vegetation) and corresponding spectra were identified and extracted from the vertices of the 2-D space plots generated by the first three components of each of the statistical analysis methods, PCA and MNF. A new dataset was generated by stacking the first three components of the PCA and MNF (in a total of six components), and a fully constrained linear SMA was implemented to map the fractional impervious surfaces. Results indicate that a promising performance has been achieved by the proposed new method with the systematic error (SE) of −3.45% and mean absolute error (MAE) of 11.52%. Comparative analysis results also show a much better performance achieved by the proposed statistical method-based SMA than the conventional SMA.

Spectral mixture analysis (SMA) is an image-processing technique used for the analysis of airborne hyperspectral remote-sensing data which consist of a large number of spectral bands, typically over 100. In this paper the possibilities are examined of using SMA for the analysis of Landsat Thematic Mapper satellite data with only six bands in the visible to shortwave-infrared wavelength. We use data from a semi-arid area in the Sanmatenga province of Burkina Faso, an area known to experience land-degradation problems. In SMA, we assume that pixels in an image consist of one or more homogeneous (uniform in character) or pure surfaces, the so called end-members. The end-members were derived from the image data on the basis of specific image characteristics. Field data was collected to describe the characteristics of these end-members in terms of land cover, soil and degradation phenomena. The end-members derived from the image analysis, although statistically pure in terms of image spectral characteristics, prove to be mixtures at a field scale. This lack of purity is explained by the nature of semi-arid areas which is more heterogeneous than that of most other areas. The SMA yielded cover percentages for the end-members from which an unsupervised classification was made. Comparison of the classification with the results of SMA shows that the latter provides information on the amount and spatial distribution of land characteristics such as land degradation.

The spectral reflectance of agricultural crops is affected significantly by sub-pixel scale spectral contributions of background soils and shadows as viewed by a remote sensing instrument. This has meant the potential of remote sensing imagery has not been fully realized for extracting biophysical information and assessing ecological stress using methods such as vegetation indices (VIs). In this paper, we address this problem explicitly using spectral mixture analysis (SMA) to quantify the area abundance of plants, soils and shadows at sub-pixel scales with the aim of improving extraction of plant biophysical and structural information from remote sensing data. Different measurement strategies were tested in the field for acquiring reference endmember spectra of crop vegetation, soil and shadows using a field spectroradiometer for a set of potato plots in western Canada. Endmember measurements included sunlit and shadowed spectra of in situ crop targets, optically thick stacks and data from excised leaves, as well as cultivated, rough and compacted soils. All possible combinations of crop, soil and shadow endmember spectra were analysed using SMA to derive sets of sub-pixel scale component fractions from radiometer spectra acquired from a boom truck over replicate plot samples with a sensor field of view of 1.05 m. Digital video image frames captured simultaneously with the radiometer data were used to determine ground proportions of crop, soil and shadow for independent validation of the SMA fractions. Endmember fractions derived from excised leaves, cultivated soil and shadowed vegetation spectra showed the best agreement with ground truth data, with differences of only ±3.3%. These sub-pixel scale fractions were used in regression analyses to predict leaf area index, biomass and plant width with an average r2 value of 0.85 from SMA shadow fraction, which was a substantial improvement over the best VI results from NDVI, NGVI and SR (average r2 = 0.53). Perspectives on SMA at different stages in the growing season and for different crop types are provided with a recommendation that further SMA research is warranted for local to regional scale agricultural crop monitoring programmes.

Independent component analysis (ICA) has shown much success in blind source separation and channel equalization. Its applications to remotely sensed images are investigated in recent years. Linear spectral mixture analysis (LSMA) has been widely used for subpixel detection and mixed pixel classification in remote sensing image processing and already studied in Chapters 3, 8–10, . It models the spectral signature of an image pixel vector as a linear mixture of spectral signatures of targets present in the image data and the target abundance fractions are assumed to be unknown, but nonrandom constants. This chapter combines these two approaches to one, called ICA-based linear spectral random mixture analysis (LSRMA), which describes an image pixel vector as a random process resulting from a random composition of multiple spectral signatures of distinct targets in the image data. It differs from LSMA in that the abundance fractions of the target signatures in LSRMA are considered to be unknown random independent signal sources. Two major advantages can be benefited from LSRMA. First, LSRMA does not require any prior knowledge of the targets used in the linear mixture model. Second and most importantly, LSRMA models each of target signatures as an independent random signal source so that the spectral variability of target signatures can be captured more effectively in a stochastic manner. The only required knowledge for LSRMA is the number of signal sources, p assumed to be present in the image data. This number is estimated by Neyman-Pearson detection theory-based eigen-thresholding methods to be investigated in Chapter 17.