Mathematics – Logic
Scientific paper
Sep 2008
adsabs.harvard.edu/cgi-bin/nph-data_query?bibcode=2008epsc.conf..646c&link_type=abstract
European Planetary Science Congress 2008, Proceedings of the conference held 21-25 September, 2008 in Münster, Germany. Online a
Mathematics
Logic
Scientific paper
Introduction: Making use of HRSC color data Mapping outcrops of clays, sulfates and ferric oxides are basis information to derive the climatic, tectonic and volcanic evolution of Mars, especially the episodes related to the presence of liquid water. The challenge is to resolve spatially the outcrops and to distinguish these components from the globally-driven deposits like the iron oxide-rich bright red dust and the basaltic dark sands. The High Resolution Stereo Camera (HRSC) onboard Mars-Express has five color filters in the visible and near infrared that are designed for visual interpretation and mapping various surface units [1]. It provides also information on the topography at scale smaller than a pixel (roughness) thanks to the different geometry of observation for each color channel. The HRSC dataset is the only one that combines global coverage, 200 m/pixel spatial resolution or better and filtering colors of light. The present abstract is a work in progress (to be submitted to Planetary and Space Science) that shows the potential and limitations of HRSC color data as visual support and as multispectral images. Various methods are described from the most simple to more complex ones in order to demonstrate how to make use of the spectra, because of the specific steps of processing they require [2-4]. The objective is to broaden the popularity of HRSC color data, as they could be used more widely by the scientific community. Results prove that imaging spectrometry and HRSC color data complement each other for mapping outcrops types. Example regions of interest HRSC is theoretically sensitive to materials with absorption features in the visible and near-infrared up to 1 μm. Therefore, oxide-rich red dust and basalts (pyroxenes) can be mapped, as well as very bright components like water ice [5, 6]. Possible detection of other materials still has to be demonstrated. We first explore regions where unusual mineralogy appears clearly from spectral data. Hematite at Aram Chaos or Terra Meridiani [7-9] is a candidate. Bright deposits have potentially spectral signatures different to the red dust in the visible: sulfates in Juventae Chasma or Aram Chaos [9, 10] and phyllosilicates in Mawrth Vallis [11] or Nili Fossae [12] are of interest. This abstract is focused on Marwth Vallis only. HRSC spectral data: geometry and color filters The spectral data are image mosaics of five broadband spectral channels centered respectively at 440, 530, 650 and 750 nm for covering the visible range of wavelengths and 970 nm for sensitivity to the electronic absorptions of minerals present in minerals (pyroxenes, olivine). The third channel (nadir image) has a typical pixel size of 12.5 m, 25 m or 50 m. The other channels have a usual pixel size of 50 m, 100 m or 200 m that determines the spatial sampling of the spectral dataset. These data are acquired by five individual cameras oriented with a specific angle to the normal to the surface (-3°, +3°, 0° (nadir), -16° and +16° respectively). Those tilts optimize the use of a single telescope for all cameras in the available room. Thus, a given spectrum results from different proportions of shade at each wavelength. Indeed, subpixel topographic slopes that are oriented toward the instrument represent a higher proportion in the signal. This implies that shade affects the shape of HRSC spectra on a different way from pixel to pixel. This contribution has to be considered when performing spectral analysis. Level-4 color images in Digital Numbers (DNs) are registered adequately and are available to the public through the HRSCview website (http://hrscview.fu-berlin.de). A linear function converts the DNs into radiance factor (I/F). Visual interpretation Color composites Red-Green-blue (RGB) color composites of DNs images contain usable geological information. Dark basaltic sands and bright red dust appear always obvious. Materials generated from interaction of liquid water, like sulfates and phyllosilicates form generally bright outcrops with complex contour lines that allow visual discrimination, even if this bright color is similar to well-illuminated bright red dust. When the surface is spectrally diverse like Marwth Vallis, contrast enhancement may be sufficient to reveal subtle color differences that correspond to different types of materials (Fig. 1a). However, those remain faint color variations as all the bands are highly correlated. Principal Component Analysis (PCA) PCA is a tool for decorrelation and noise removal that maximizes color unit differences. On Marwth Vallis, PCA highlights the diversity of the surface on a spectacular way (Fig. 1b). Those images may be compared to the maps of mineral composition obtained by [11] from spectral analysis imaging spectrometer data. Part of the information in Fig. 1b is likely related to surface roughness because of the complex geometry of observation of the instrument. Furthermore, only an extremely clear atmosphere and low-compressed datasets allow obtaining such sharp results. Consequently, the meaning of the colors varies from image-to-image and is qualitative only. More quantitative and comparable results require spectral analysis, either to remove or to normalize atmospheric and geometric effects. Spectral analysis on HRSC data For this application the surface units to be distinguished have to possess linear independent color vectors in the five-dimensional color space of HRSC data. It has been shown by [2-5] that on the global scale, only four spectral endmembers representing red, iron oxide-rich material, dark, basaltic material, and ice plus a shade component containing effects of observation and illumination geometry, are sufficient to explain most of the colors present in HRSC color imagery. We assess this at our test areas contain a maximum of surface mineralogy diversity by applying refined methods to model (and remove) the shade contribution in order to test if a further surface component can be unambiguously detected in the HRSC color dataset. Error! Reference source not found.a shows that Spectral Mixing Analysis (SMA) performed by the Multiple-Endmember Linear Spectral Unmixing Model (MELSUM) [9] is able to separate bright red dust and bright outcrops known as hydrous materials. Root-Mean Square (RMS) model residuals mostly contain effects due to topography. Perspectives We will continue to investigate HRSC color data to map surface units and consider material diversity, atmospheric opacity, illumination and observation geometry, and calibration. Coming results will determine in which cases visual interpretation is sufficient, how spectral analysis can be performed to map surface units, and how take the advantage of imaging spectrometry. References [1] Neukum G. et al. (2004), ESA-SP 1240. [2] Combe J.-Ph. et al. (2007) 38th LPSC 2367. [3] Combe J.-Ph. et al. (2008) 39th LPSC 2381. [4] Wendt L. et al. (2008) 39th LPSC 1242. [5] McCord T. B. et al. (2007) JGR 112. [6] McCord T. B. et al. (2006) LPSC 1757. [7] Christensen P. et al. (2001), JGR 106 E10. [8] Glotch, T. D. et al. (2005), JGR, 110, E9. [9] Combe J.-Ph. et al. (2008), PSS 56. [10] Gendrin A. et al.(2005) Science 307. [11] Loizeau D. et al. (2007) JGR 112. [12] Mangold, N, et al. (2007), JGR, 112, E08S04. Acknowledgements First and third authors acknowledge NASA for contract with the Mars-Express mission. Second and Fourth authors acknowledge the German Space Agency (DLR Bonn) for their financial support of this study.
Combe J.-Ph.
Gerhard Neukum
McCord Th. B.
Wendt Lorenz
No associations
LandOfFree
Detailed mapping of surface units on Mars with HRSC color data does not yet have a rating. At this time, there are no reviews or comments for this scientific paper.
If you have personal experience with Detailed mapping of surface units on Mars with HRSC color data, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Detailed mapping of surface units on Mars with HRSC color data will most certainly appreciate the feedback.
Profile ID: LFWR-SCP-O-1793969