Evaluations of classification and spectral unmixing algorithms using ground based satellite imaging

Abstract

Abundances of material components in objects are usually computed using techniques such as linear spectral unmixing on individual pixels captured on hyperspectral imaging devices. However, algorithms such as unmixing have many flaws, some due to implementation, and others due to improper choices of the spectral library used in the unmixing (as well as classification). There may exist other methods for extraction of this hyperspectral abundance information. We propose the development of spatial ground truth data from which various unmixing algorithm analyses can be evaluated. This may be done by implementing a three-dimensional hyperpspectral discrete wavelet transform (HSDWT) with a low-complexity lifting method using the Haar basis. Spectral unmixing, or similar algorithms can then be evaluated, and their effectiveness can be measured by how well or poorly the spatial and spectral characteristics of the target are reproduced at full resolution (which becomes single object classification by pixel).

Hyperspectral feature classification with alternate wavelet transform representations

“…Thus the wavelength range of the data is from the visible to the shortwave-infrared (SWIR). The spacecraft was comprised of only four material spectra (Figure 7) used in the simulation: aluminum, mylar, solar cells, and white paint, as illustrated in Figures 6a-e. Data such as in Figures 5, 6 and 7 would typically be used for a computational testbed as described in earlier papers 3, 8 . Actual implementation of this enables tasks such as spectral library manipulation, adding noise and other image degradation, wavelet transforms, and feature extraction algorithms.…”

Accelerating computed tomographic imaging spectrometer reconstruction using a parallel algorithm exploiting spatial shift-invariance

“…One instrument for snapshot hyperspectral imaging is CTIS which is robust, simple, fast, and has military successes in missile defense, ISR and target/threat characterization/detection. 1,2,9,16,17 Even though CTIS captures raw images in milliseconds, 1,11,13,[18][19][20][21] post processing the imagery requires enormous computational resources, typically taking minutes to hours. 18,22 Since potential applications for CTIS are embedded realtime or near-realtime systems, a better reconstruction algorithm is needed.…”

Tensor methods for hyperspectral data analysis: a space object material identification study

“…Here we consider more recent hyperspectral image data of the type collected by the AEOS ASIS system on Maui. Work related in various ways to ours can be found, e.g., in papers by Blake et al [14][15][16], Hege et al [29], and Scholl et al [30].…”