Rama Rao Nidamanuri

Normalized Spectral Similarity Score (

We present a new spectral library search algorithm, referred to as normalized spectral similarity score (NS3), for improved accuracy in airborne hyperspectral image classification. The proposed library search algorithm combines the relative merits of spectral angle and amplitude differences inherent in a hyperspectral image and reference library reflectance spectra. Various spectral libraries constructed from the field reflectance spectra collected during two successive growing seasons were used for classification of a historical HyMAP hyperspectral image for crop classification by spectral library search approach. The performance of the proposed method was compared with existed spectral library search methods, i.e., spectral angle mapper (SAM), spectral correlation mapper (SCM), spectral information divergence (SID), and the classical maximum likelihood classifier (MLC). The best classification accuracy obtained from the proposed NS3 library search method (74.71%) was significantly lower than that of the MLC supervised classification (85.44%). However, a comparative analysis of the classification accuracy indicates the enhanced performance of the proposed NS3 method for transferring a spectral library for HyMAP image classification, because the classification accuracies of the other library search methods tested were considerably lower (MC (61,87%), SAM (54,10%), SCM (52,51%), and SID (34,30 %)). Furthermore, various factors that influence the performance of spectral library search method for hyperspectral image classification are discussed.

{\hbox{NS}}^{3}

Accurate generation of a land cover map using hyperspectral data is an important application of remote sensing. Multiple classifier system (MCS) is an effective tool for hyperspectral image classification. However, most of the research in MCS addressed the problem of classifier combination, while the potential of selecting classifiers dynamically is least explored for hyperspectral image classification. The goal of this paper is to assess the potential of dynamic classifier selection/dynamic ensemble selection (DCS/DES) for classification of hyperspectral images, which consists in selecting the best (subset of) optimal classifier(s) relative to each input pixel by exploiting the local information content of the image pixel. In order to have an accurate as well as computationally fast DCS/DES, we proposed a new DCS/DES framework based on extreme learning machine (ELM) regression and a new spectral-spatial classification model, which incorporates the spatial contextual information by using the Markov random field (MRF) with the proposed DES method. The proposed classification framework can be considered as a unified model to exploit the full spectral and spatial information. Classification experiments carried out on two different airborne hyperspectral images demonstrate that the proposed method yields a significant increase in the accuracy when compared to the state-of-the-art approaches.

) as an Efficient Spectral Library Searching Method for Hyperspectral Image Classification

Use of spectral library searching as an automated method of analysing hyperspectral remote sensing data for material mapping is gaining prominence, especially in the mineral mapping domain. The possibility and reliability of material identification by the spectral library search approach depends on the spectral representation, characterized by spectral resolution (or sampling interval) and comparison metric used. We present a method referred to as Relative Search Performance (RSP) for an evaluation of various spectral representations and comparison metrics for designing an optimal library search system for material mapping. The proposed method works on the basis of tracking the changes in the spectral matching ranks of entries in the hit lists of spectral library searches for various spectral representations and comparison metrics relative to a chosen standard. The method has been tested for the comparison of the search performance of various discrete spectral sampling intervals and popular comparison metrics using the USGS Spectral Library. Results indicate that this approach can be used for the selection of optimal spectral representation and/or for selecting a comparison metric appropriate for a particular material mapping application by the reflectance spectral library search.

Dynamic Ensemble Selection Approach for Hyperspectral Image Classification With Joint Spectral and Spatial Information

Imaging spectroscopy is an emerging and versatile technique that finds applications in diverse fields concerned with remote identification, discrimination and mapping of materials. The large amount of spectral data produced by hyperspectral imaging necessitates the development of automated techniques that convert imagery directly into thematic maps. Spectral library search method, a method of choice for organic compound identification by the mass spectroscopy, has caught the attention of researchers as one of the appropriate methods for an efficient exploitation of high quality spectral data available from the hyperspectral imaging systems. Given the apparent increase in the number of papers appearing on the subject as well as the variety of methods proposed, it is reasonable to say that the field of automated interpretation of reflectance spectral data has passed its infancy now gaining important space in the scientific community. We present an overall view of the literature relevant to the development of library search method, the various search algorithms and systems available in the purview for developing an automated hyperspectral data analysis system for material identification.

A method for selecting optimal spectral resolution and comparison metric for material mapping by spectral library search

ABSTRACT Three-dimensional (3D) point cloud labelling of airborne lidar (light detection and ranging) data has promising applications in urban city modelling. Automatic and efficient methods for semantic labelling of airborne urban point cloud data with multiple classes still remains a challenge. We propose a novel 3D object-based classification framework for labelling urban lidar point cloud using a computer vision technique, supervoxels. The supervoxel approach is promising for representing dense lidar point cloud in a compact manner for 3D segmentation and for improving the computational efficiency. Initially, supervoxels are generated by over-segmenting the coloured point cloud using the voxel-based cloud connectivity algorithm in the geometric space. The local connectivity established between supervoxels has been used to produce meaningful and realistic objects (segments). The segments are classified by different machine learning techniques based on several spectral and geometric features extracted from the segments. All the points within a labelled segment are assigned the same segment label. Furthermore, the effect of different feature vectors and varying point density on the classification accuracy has been studied. Results indicate an accurate labelling of points in realistic 3D space conforming to the boundaries of objects. An overall classification accuracy of is achieved by the proposed method. The labelled 3D points can be used directly for the reconstruction of buildings and other man-made objects.

Spectral material mapping using hyperspectral imagery: a review of spectral matching and library search methods

The urban land cover mapping and automated extraction of building boundaries is a crucial step in generating three-dimensional city models. This study proposes an object-based point cloud labelling technique to semantically label light detection and ranging (LiDAR) data captured over an urban scene. Spectral data from multispectral images are also used to complement the geometrical information from LiDAR data. Initial object primitives are created using a modified colour-based region growing technique. Multiple classifier system is then applied on the features extracted from the segments for classification and also for reducing the subjectivity involved in the selection of classifier and improving the precision of the results. The proposed methodology produces two outputs: (i) urban land cover classes and (ii) buildings masks which are further reconstructed and vectorized into three-dimensional buildings footprints. Experiments carried out on three airborne LiDAR datasets show that the proposed technique successfully discriminates urban land covers and detect urban buildings.

A supervoxel-based spectro-spatial approach for 3D urban point cloud labelling

Driven by significant technological developments in the hyperspectral imaging, material mapping using reference spectra has received renewed interest of the remote sensing community. The applicability of reference spectral signatures in image classification depends mainly on the material type and its spectral signature behaviour. Identification and spectral characterization of materials which exhibit unique spectral behaviour is the first step in this approach. Consequently there have been active researches for the identification of surface materials which exhibit unique spectral signatures. The uniqueness of reflectance signature of winter rape relative to its co-occurring crop species was reported in this study. Reflectance spectral libraries constructed from field spectral reflectance measurements collected over five agricultural crops (alfalfa, winter barley, winter rape, winter rye, and winter wheat) during four subsequent growing seasons were classified by the linear discriminant analysis (LDA). Further, the reference field spectral database was used for the spectral feature fitting and classification of a historical HyMAP airborne hyperspectral imagery acquired at a separate site, by spectral library search. Results indicate the existence of a meaningful spectral matching between image and field spectra for winter rape and demonstrate the potential for transferring spectral library for hyperspectral image classification. The observed consistency in the discrimination of winter rape demonstrates experimentally the fundamental principle of remote sensing which suggests the theoretical existence of unique spectral signatures for materials which can be incorporated as reference spectral signatures for hyperspectral image classification.

Object-oriented semantic labelling of spectral–spatial LiDAR point cloud for urban land cover classification and buildings detection

A study was carried out examining the effects of conservation policy on the ecosystem and livelihoods of local people on the Rajiv Gandhi National Park, located in one of the global biodiversity hotspots, the Western Ghats, India. Results show that less than 5% of the people are in favor of the policies while a staggering 94% of the people are strongly against the policies. The remaining 1% of the total respondents are found neutral with regard to conservation policies. Several reasons viz., ban on agriculture, restriction on livestock rearing and grazing, ban on non-timber forest collection, exclusion of local and indigenous communities in conservation programs and tourism activities are found to be responsible for negative attitude towards the national park. Apart from limiting the local livelihood options, the anticipated ecological consequences are not encouraging as exotic species are dominating the vegetation dynamics of the area—replacing many native plant species. The fact is that the inextricable link between nature and society needs an integrated science-policy research approach for biodiversity conservation in the hotspots, particularly in the developing countries where human and ecosystem interactions are much more complex and closely interwoven with each other.

Understanding the Unique Spectral Signature of Winter Rape

Spectral library search is emerging as a viable approach for material identification and mapping by reusing spectral knowledge gained from hyperspectral remote sensing across space and time. The potential of retrieving meaningful spectral material identifications in the presence of reflectance of spectra of various material types and with various similarity metrics has been assessed in this study. Test reflectance spectra of various vegetation, minerals, soils and urban material types are identified by searching through the composite reflectance spectral library obtained by combining various institutional reflectance spectral libraries. The accuracy of material identifications under various conditions: (i) in the presence of identical, similar and dissimilar spectra; (ii) in the presence of only identical and dissimilar spectra; and (iii) in the presence of only dissimilar spectra has been assessed with several similarity metrics. Results indicate the possibility of obtaining 100% accurate material identifications by library search if the spectral library contains identical spectra. However, the presence of a large number of similar spectra, despite the presence of identical spectra, is found to increase false positives, thereby reducing the accuracy of retrievals to 82% at best. Further, the accuracy of material identifications in the presence of similar spectra is similarity metric-dependent and varied from about 52% (obtained from Binary Encoding) to 82% (obtained from Normalized Spectral Similarity Score). Overall, results support the possibility of using independent reflectance spectral libraries for material identification while calling for robust spectral similarity metrics.

Ecological and socioeconomic impacts of conservation policies in biodiversity hotspots: a case study from Rajiv Gandhi National Park, India

Image classification for material mapping using independent training data is emerging as an automated method for hyperspectral image analysis. Possibility of using independent training data for image classification depends upon material type and its spectral behaviour. Identification and spectral discrimination of materials which exhibit characteristic spectral behaviour are critical for developing hyperspectral material detection and mapping methods. We identify and evaluate characteristic reflectance signature of winter rape relative to its co-occurring crops from a hyperspectral image classification perspective. Spectrallibraries developed using field reflectance measurements of agricultural crops: alfalfa, winter barley, winter rape, winter rye and winter wheat collected during four growing seasons are searched through for the classification of a HyMap image acquired for a separate site by spectral angle mapper and spectralfeature fitting methods. Results indicate the existence of a characteristic spectral signature for winter rape and meaningful matching between image and field spectra, which can be used for automatic mapping of winter rape by hyperspectral imaging.