Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.stsci.edu/~mutchler/documents/ACM_2008_poster.pdf Дата изменения: Mon Jul 7 18:41:44 2008 Дата индексирования: Sat Sep 6 00:49:40 2008 Кодировка: Поисковые слова: ceres

S tr a te g ie s fo r im a g in g C e r e s a n d V e s ta w ith th e H u b b le S p a c e T e le s c o p e
Max Mutchler (Space Telescope Science Institute), Lucy McFadden, Jian-Yang Li (U. Maryland), Peter Thomas (Cornell U.), Joel Parker, Eliot Young (Southwest Research Institute), Chris Russell, Britney Schmidt (UCLA), Mark Sykes (Planetary Science Institute)

Figure 2: Drizzled ACS color composite (F330W + F555W) image of Ceres

Figure 3: Deconvolved W FPC2 color composite (F439W + F673N) image of Vesta, with graphic of the Dawn spacecraft (encounter in 2011)

geometric distortion which ------------------------------------The --be-minimal-since-Vesta-a-slight --placed-near-the-PC ---was-not-removed ----effects ---but-the-effect--------- WFPC2 camera produces always --- -- - -- chip center, where-- from the-images, smallest. --- - --- distortion - -- are - - -In should - ---- -- --- is --preparation for deconvolution, the images were cropped to a small area centered on Vesta, and a simple cosmic ray rejection process was performed on individual images (no image combination) with the disk of Vesta masked to prevent any rejections there. The d econvolution is very sensitive to any CCD defects (e.g. hot pixels) or unrejected cosmic rays, and this may explain some artifacts seen in the output images. Also, MEM deconvolutions of earlier WFPC2 images of Vesta in 1994 and 1997 [6,7] exhibited a ringing effect on the bright sunlit limb of Vesta, and this effect is likely present again. We continue to experiment with independent deconvolution methods, which should provide further leverage to help discern real features from artifacts.

Introduction: The Hubble Space Telescope (HST, Figure 1) was used to conduct highresolution imaging of Vesta and Ceres in support of the Dawn mission, which will encounter them in 2011 and 2015, respectively. We describe two quite different observational and data processing strategies used to extract as much spatial information as possible for the shape and surface features of these two objects.

Deconvolving Vesta: Planetary Camera (PC) images of V esta w ere deconvolved using the Maximum Entropy Method (MEM), as implemented in the IRAF/STSDAS r estore package at STScI [4]. The TinyTIM package [5] was used to make PSFs for each filter. All PSFs were s ubsampled by a factor of four, and output images were produced at one-quarter of the input detector scale, or 0.0114 arcsec/pixel (Figure 3).

Hubble observations: Imaging of Ceres was conducted with the ACS High Resolution Channel (HRC) in 2004 (HST program 9748). Full rotation imaging allowed for analyses of its shape and construction of a global albedo map [1,2]. This program also produced subsampled data for three phase angles 120 degrees apart, with filters F330W and F555W. The 9.1 hour rotational period of Ceres is slow enough to execute a 4-exposure half-pixel dither "box" pointing pattern. This creates a dataset which can be drizzle-combined to enhance the spatial resolution. Following the failure of ACS in January 2007, the May 2007 observations of Vesta had to be conducted with WFPC2 (HST program 10799). Vesta's 5.3 hour rotational period is too fast to execute a subsampling dither pattern at each phase. So a series of single images were obtained which sampled many more phase angles with filters F439W, F673N, F953N, F1042M, and deconvolution methods were applied to each individual image. Drizzling Ceres: The four ACS/HRC exposures for each filter were carefully registered using a cross-correlation method which utilizes the available surface features to align the images to within a small fraction of a pixel. Then they were distortioncorrected, combined, and cleaned of cosmic rays and detector artifacts using MultiDrizzle [3]. The subsampled data was drizzled to an output scale of 0.015 arcsec/pixel, or 40% smaller than the input detector pixels (Figure 2).

Figure 1: Hubble Space Telescope

Conclusion: The drizzled and deconvolved output pixel scales were somewhat arbitrarily chosen to extract as much spatial information as possible from the data. The actual improvement in resolution is very difficult to quantify, it varies by wavelength, and some rotational blurring is certainly involved. But the resulting images clearly reveal and define surface features, which exhibit rotational motion, and allows for their physical interpretation. Albedo features of Ceres can be seen in Figure 2 that have been calibrated in photometric maps to a range of 14% relative to the mean single particle scattering albedo (SSA) of Ceres (0.073 at V-band) and color ratios vary by 12% [2]. Albedo features of Vesta seen in Figure 3 range from -17 to +17% of the global average SSA (0.42 at V-band) at all four wavelengths. Color variations can be related to possible compositional variations [8]. References: [1] Thomas, P. et al. (2005) N ature, 437, 224. [2] Li, J. et al. (2006) I carus, 182, 143. [3] Koekemoer, A. (2002) HST Calibration Workshop, 337. [4] Wu, N. (1995) ADASS IV, ASP Conf. Series Vol. 77. [5] K rist, J. & Hook, R. (2004) TinyTIM Users Guide [6] Zellner, B., et al., (1997) I carus, 128, 83. [7] Thomas, P. et al., (1997) Science, 277, 1492, [8] Li et al., (2008) ACM