Fast searching of spectral library database using variable interval spectral average method

Abstract

With ever-increasing number of spectral channels from space-borne hyperspectral instruments, demand on approaches for fast search schemes for matching hyperspectral pixel vector with standard spectral library database has increased proportionately. The present-day methods are tedious and time consuming to meet the above task. We propose a fast matching scheme based on bivariate short-interval local variance that can be used to capture the essence of reference materials in the spectral library. The variance of each selected window is computed across the spectral curve data and the peak variance above a threshold is taken as a spike. The position and linewidth of the spikes are shown to carry unique signatures of the given material spectral data, which can be stored and used as matching criteria. The choice of appropriate threshold is important; it has been found that the mean value of background variance signal could be used as the threshold value. The proposed method was successfully applied to identify some samples of the AVIRIS hyperspectral imagery to the standard JPL spectral library database.

Spectral matching approaches in hyperspectral image processing

“…Here, differential weights are assigned to the wavelength ranges of the spectrum based on their importance for matching. Kumar et al (2006) introduced the Variable Interval Spectral Average (VISA) method, where variance is computed for each region in the spectrum. When the variance exceeds a threshold, the presence of a spike is confirmed.…”

Hyperspectral image classification by a variable interval spectral average and spectral curve matching combined algorithm

“…The VISA method is a variant of time average method proposed for locating coherent packets in a turbid liquid flow (Hudgins and Kaspersen, 1999;Kumar et al, 2006). In the VISA method, we basically compute significant variations, if any, present in the signal by estimating 'local' or 'short-interval' variances.…”

“…Secondly, treating these two parameters as independent measures, the pixel is assigned to a class label independently for which each of these values were at its minimum. Finally, the pixel would then be assigned to that class label for which both the class labels are same; or else, the pixel is declared as 'unclassified', thus giving both these measures equal bias in judging the final decision (Kumar et al, 2006).…”

“…This is done with combination of two schemes-variable interval spectral average (VISA) method (Kumar et al, 2006) and spectral curve matching (SCM) method (Van der Meer, 2000). More specifically, the VISA method would help to detect 'transient' signal variations of spectral data at different scales of windows, while the SCM method would help to find a match between these transient features of the pixel and one of spectra from standard spectral library by least square fitting.…”

A Real-Time Infrared Ultra-Spectral Signature Classification Method via Spatial Pyramid Matching

“…There is a continuous demand for reducing time consumption and improving the accuracy of classification algorithms [ 18 , 19 ]. The spatial pyramid matching (SPM) method [ 20 ] has been demonstrated as an excellent feature extraction method and is widely used in image feature extraction [ 21 , 22 ] and image classification [ 23 , 24 ].…”