PDS_VERSION_ID = PDS3 LABEL_REVISION_NOTE = "2006-06-21, RL 2007-03-15 2007-04-10, BSword 2007-04-26, RL 2007-05-15, BSword 2007-09-20, RL" RECORD_TYPE = STREAM OBJECT = DATA_SET DATA_SET_ID = "MRO-M-HIRISE-3-RDR-V1.0" OBJECT = DATA_SET_INFORMATION DATA_SET_NAME = "MRO MARS HIGH RESOLUTION IMAGING SCIENCE EXPERIMENT RDR V1.0" DATA_SET_COLLECTION_MEMBER_FLG = "N" DATA_OBJECT_TYPE = "IMAGE" START_TIME = 2006-09-29T15:16:33.333 STOP_TIME = "NULL" DATA_SET_RELEASE_DATE = "NULL" PRODUCER_FULL_NAME = "ALFRED MCEWEN" DETAILED_CATALOG_FLAG = "N" ARCHIVE_STATUS = "IN QUEUE" CITATION_DESC = "McEwen, A., Mars Reconnaissance Orbiter High Resolution Imaging Science Experiment, Reduced Data Record, MRO-M-HIRISE-3-RDR-V1.0, NASA Planetary Data System, 2007." DATA_SET_TERSE_DESC = "Radiometically-corrected, geometrically-mapped images for HiRISE (High Resolution Imaging Science Experiment)." ABSTRACT_DESC = "This dataset includes reduced data records from the HiRISE instrument on MRO." DATA_SET_DESC = " Data Set Overview ================= The High Resolution Imaging Science Experiment (HiRISE) is one of the remote sensing instruments on the Mars Reconnaissance Orbiter (MRO) spacecraft that acquires orbital observations of the Martian surface during a two earth-year primary mapping phase. MRO, successfully launched in August 2005, arrived at Mars in March 2006. Following orbit insertion the spacecraft went into an aerobraking period to achieve a 250 x 315 kilometer near-polar orbit suitable for the Primary Science Phase (PSP) mapping that started in November 2006. Since the start of PSP HiRISE has been continuously operating acquiring 10-20 observations per day. The HiRISE team is responsible for maintaining an updated dataset of the best version of its science data until meaningful changes in data calibration no longer occur and to release data in an appropriate manner for public access including their final deposition to NASA's Planetary Data System (PDS). In carrying out these responsibilities, the HiRISE team creates two types of standard data products: 1) Experiment Data Record (EDR) products and 2) Reduced Data Record (RDR) Products. This document describes the RDR standard products. These Reduced Data Record (RDR) products are radiometrically- corrected images resampled to a standard map projection. They are formatted and organized according to the standards of the PDS. The RDR image is stored in the JPEG2000 format recently accepted by the PDS. The JPEG2000 images are accompanied by a PDS detached label providing supporting information about the observation. The HiRISE RDR products comply with the PDS standards for file formats and labels, specifically using the PDS image object definition. The RDR image files, formatted according to the JPEG2000 standard, use 'JP2' as their filename extension. They are accompanied by PDS labels; files that have the same name as the image data file but use 'LBL' for their filename extension. The label file provides image data characterization and science metadata information about the observation. Additionally, the ancillary data files that accompany the RDR products and the archive volume structure are in conformance with PDS standards. RDR image data are stored in the JPEG2000 ISO/IEC Part 1 standard format (http://www.jpeg.org/jpeg2000/), which was accepted by the PDS standard in October 2005. The JPEG2000 standard offers benefits distinctly advantageous for storage and access to very large images. With HiRISE RDR products reaching sizes exceeding 30,000 x 70,000 pixels the use of JPEG2000 was recognized as a suitable solution for the storage and distribution of these data products. Advantages include excellent compression performance, multiple resolution levels from a single image data set, progressive decompression quality layers, lossless and lossy compression (HiRISE RDR products use lossless compression per the PDS Standard), pixel datum precision up to 38 bits, multiple image components (or bands), and selective image area access. These features are achieved by the use of a sophisticated image coding system based on discrete wavelet transforms (DWT) combined with other coding techniques to generate a JPEG2000 codestream that can be rendered to image pixel rasters using inverse transform algorithms. Processing ========== Science data from the MRO payload experiments are packetized on the spacecraft, transmitted to Earth through the Deep Space Network, and sent to the Jet Propulsion Laboratory (JPL) through ground communications. The JPL Multi-Mission Operations Facility converts the packetized data back to the original science data format as produced by the instruments. For HiRISE observations, a raw science product is created for each CCD/Channel involved in the observation. The data are stored at JPL's Raw Science Data Server (RSDS) for access by the science teams. At the HiRISE Operations center (HiROC) we have developed a ground data system that provides automated methods for retrieving and processing our images. Science data are automatically retrieved from the RSDS and passed to a series of pipeline procedures managed under the Conductor environment (http://pirl.lpl.arizona.edu/software/Conductor.shtml). The pipelines generate intermediate products used for image evaluation and standard data products for science analysis. Additionally, the pipelines populate HiRISE-catalog EDR and RDR product tables and observation geometry tables with relevant metadata. The product and geometry tables are used to generate index tables provided as part of the product distribution to the PDS. The HiRISE team verifies observations were properly acquired and science objectives achieved. Missed targets or observations with poor viewing conditions are flagged for reacquisition at a later time. HiROC uses the Integrated Software for Imagers and Spectrometers (ISIS) system in the pipeline processing. ISIS contains a wide range of tools including radiometric calibration and cartographic processing procedures. The ISIS components applicable to HiRISE include radiometric calibration, map projection transformation, image mosaicking, camera pointing correction, and general image enhancement, display, and analysis tools. For more information on this freely available image analysis package, see the ISIS web site (http:// isis.astrogeology.usgs.gov/). HiROC's pipeline processing starts when the FEI_WatchDog procedure, responsible for periodically monitoring the RSDS, determines a raw science data file is ready to be downloaded from the RSDS and passes the name of the product to the HiDog pipeline (Downlink Organizer). The HiDog pipeline retrieves the data product then submits the file name to the EDRgen pipeline (EDR generator) for creating the EDR product and populating the EDR product table with metadata about the product. The HiRISE Observation software used to produce the EDR products is described in detail at http://pirlwww.lpl.arizona.edu/software/HiRISE/ and can be obtained on request. The EDR_Stats pipeline creates an ISIS file and generates image statistics that are placed in the EDR products table. The HiCal pipeline performs the radiometric correction (see Section 4.1.1), and generates a browse and thumbnail image of the EDR product to be used by the HiRISE and PDS Imaging Node data-distribution web services. When the HiCal pipeline determines two channel files of a CCD have been calibrated the file names are passed on to the HiStitch pipeline for creating an intermediate CCD image file. HiStitch additionally adjusts the channels to radiometrically match at the seam where the two channels come together. When the HiStitch pipeline determines all of the observation's CCDs have been stitched together for a filter set then the HiccdStitch pipeline is invoked to create an intermediate product with the CCDs stitched together to form a single image for the color filter. HiccdStitch additionally adjusts the CCD images to radiometrically match the overlapping areas of adjacent CCDs. Finally HiccdStitch creates an intermediate JPEG image for evaluation by the HiRISE team. At this point in the processing the EDR-validation step occurs. To continue the pipeline processing the reconstructed SPK and CK SPICE kernels, providing information about the observing viewing geometry and spacecraft ephemerides, need to be retrieved from the NAIF Node though HiROC's HiSPICE subsystem. Reconstructed SPICE kernels are generally provided to the MRO science teams one to two weeks after the observation was acquired. The HiGeomInit pipeline extracts the spacecraft ephemeris kernel (SPK) and C-matrix pointing kernel (CK) data from the SPICE files and transfers the data to the CCD image files for geometry processing by the pipelines that follow. Additionally HiGeomInit populates the observation geometry table with viewing geometry and coordinate metadata. The RedGeom and ColorGeom pipelines perform geometric processing on individual CCD images. The RedMosaic and ColorMosaic pipelines mosaic the projected CCD images to form an observational image. Additionally these pipelines create RDR-product browse and thumbnail jpeg images for HiRISE and PDS Imaging Node web-based distribution services. The last pipeline step, RDRgen, creates the JPEG2000 formatted JP2 image file accompanied by a PDS detached label file and populates the RDR product table with information about the product. The software for converting a raw PDS file with an attached label to PDS label and JP2 files is available on request. Following a final validation step the EDR and RDR products are releasable to the PDS and science community. Radiometric Calibration Processing ---------------------------------- The radiometric calibration-correction procedure is described here at a high level. A detailed description will be provided in a future HiRISE calibration paper. The radiometric calibration correction is performed on each individual HiRISE channel file (EDR) correcting for instrument offset, dark current, gain, then converting to I/F reflectance. The first step in the calibration, carried out by the ISIS hiclean program, corrects for instrument dark current and offset. The hiclean program uses the ancillary calibration data (dark and mask pixels that accompany the science data to compute corrections in both the column (sample) and row (line) directions. The mask pixels, positioned at the start of the instrument output, provide dark current information for each column. The dark pixels, positioned at the end of each image row, capture the time dependent dark current and offset instrument drift. The ISIS hical program then applies an intra-channel B0 (additive dark current matrix) and A0 (multiplicative gain matrix) correction for each column in the image array. The hical then converts the pixel values to I/F (intensity/flux, I/F = 1 for a 100% ideal lambertian reflector viewed normal to the surface) as described below: For: H = dark current and offset corrected image, output of hiclean B0 = intra-channel dark current correction (TDI & BIN dependent) A0 = intra-channel gain correction (TDI and BIN dependent) G = global gain correction, normalizes CCD/channels L = observation line time I = I/F conversion factor at Sun-Target distance of 1.5 AU AU = Mars-to-Sun distance (AU) at time of observation Z = radiometrically corrected image in I/F units The correction is: Z = ([H-(B0_L)]/L)_A0_G*I*(1.5/AU)2 Instrument instabilities result in radiometric mismatches requiring additional corrections for the varying column-to-column, channel-to- channel, and CCD-to-CCD sensitivities. Residual column-to-column variations are corrected by first computing the mean value for each column in an image array. The mean-value one-dimensional array is then high-pass filtered to eliminate low-frequency information due to scene content. The result of the high pass filter is then subtracted from the image array. CCD channels are adjusted to radiometrically match at the seam where the two channels come together (performed by the HiStitch pipeline). The CCDs are then radiometrically matched one to the other by matching the overlapping areas of adjacent CCDs ( performed in the HiccdStitch pipeline). Geometry Processing ------------------- There will typically be two RDR standard products per observation: a single-color RDR product built from the operating red-filter CCDs, and 0a three-color RDR product if the blue-green and near-infrared CCDs were additionally operating. In rare cases a two-color RDR product will be created if only two color filters were commanded. The geometric processing corrects for the optical distortion and projects the observation from spacecraft viewing orientation to a map coordinate system. For RDR products the Equirectangular or Polar Stereographic projections are used. The geometry processing, carried out by ISIS program cam2map, uses cubic convolution resampling. Geometry processing employs the NAIF toolkit (http://naif.jpl.nasa.gov) and uses reconstructed SPICE kernels generated by the MRO project. The geometry processing uses the MOLA Digital Terrain Model to improve the camera pointing intercept position on the Martian surface. In the geometric processing, individual channel images are stitched together to form CCD images using the ISIS program histitch. The spiceinit program searches through the available NAIF kernel set and applies planet and spacecraft ephemeris data to establish geometric properties of each CCD image CCD images are then individually map projected with camp2map and mosaicked together using himos forming an image of the entire observation. Resulting image maps vary in size depending on the number of CCDs commanded, number of lines acquired and the binning mode of the images. RDR products can be very large, at times exceeding 30,000 x 70,000 pixels. Observations with mixed binning modes are resampled to the same pixel scale depending on the minimum binning used in the observation. For the Transition Orbit Phase (TRA) and Primary Science Phase (PSP), observations with unbinned imaging are uniformly mapped to a constant 0.25 m/pixel resolution (0.50 m/pixel for minimum binning 2 and 1.0 for binning 4). The Aerobraking phase images (AEB) are mapped to a pixel scale depending on the spacecraft altitude and the minimum binning. For three-color imaging additional processing steps are required to create the three-color RDR products. Spacecraft jitter exists at a higher frequency than can be captured by the reconstructed pointing kernels. The color CCDs see a point on the planet surface separated by ~120 milliseconds resulting in jitter-induced pointing misalignments not captured by the reconstructed pointing matrix. To minimize color misregistration the blue-green and near-infrared filters are spatially matched through a two-step empirical approach. In the first step the blue-green and near-infrared filter CCD images (usually acquired at a higher binning level than the red-filter CCDs) are scaled to match the binning of the red-filter CCD imaging. Next, an ISIS program, hijitreg, applies an image coregistration process constructing a control network that spatially maps the blue-green and near-infrared filter imaging to the red filter imaging. The control network is provided to the second step responsible for resampling the blue-green and near-infrared filter images to the red. The program, slither, performs a one-dimensional cubic-spline transformation performing translational shifts on otherwise undistorted individual lines. Once the blue-green and near-infrared images are spatially registered to match the red imaging, the images go through the same geometric processing to create a map-projected image as described above. The BANDWIDTH keyword in the PDS labels identifies the color filters used in the image product. The CENTER_FILTER_WAVELENGTH and BANDWIDTH keywords provide information about the spectral range of each filter The storage order for the the three-color products is Near Infrared (band 1), Red (band 2), and blue-gree (band 3). Data ==== The HiRISE EDR products act as the input source to the RDR processing. Ancillary Data ============== Generation of the RDR products relies on the MRO-project deliveries for the Spacecraft Ephemeris (SPK kernels) and MRO spacecraft pointing files (CK kernels) Coordinate System & Cartographic Standards ========================================== The HiRISE RDR products are compatible with the cartographic standards and mapping conventions used by the MRO CRISM map products. When spatially registering map products produced by the two instrument teams only a translation and scale change are required. The coordinate system used is planetocentric latitude and east positive longitude direction. The planetocentric latitude is the angle from the equator to a point on the surface of an oblate planet. The longitude increases from west to east (left to right). The planetary constants used in the camera model to produce the HiRISE RDR products are obtained from the NAIF SPICE planetary constants kernel pck0008.tpc. The Mars constants of particular importance are the right ascension and declination of the pole, the prime meridian, rotation rate, and radii. The constants used are: BODY499_POLE_RA = ( 317.68143 -0.1061 0. ) BODY499_POLE_DEC = ( 52.88650 -0.0609 0. ) BODY499_PM = ( 176.630 350.89198226 0. ) BODY499_RADII = ( 3396.19 3396.19 3376.20 ) Additionally, the SPICE kernel de405.bsp was used for the ephemeris data for Mars. Two map projections are used in the HiRISE RDR products: Equirectangular and Polar Stereographic." CONFIDENCE_LEVEL_NOTE = " Confidence Level Overview ========================= Known problems are TBD. Review ===== This archival data set has been examined by a peer review panel prior to its acceptance by the Planetary Data System (PDS). The peer review has been carried out in accordance with PDS procedures. Data Coverage and Quality ========================= RDR products are the permanent record of the HiRISE observations that are radiometrically corrected an map projected. The observational coverage on the Martian surface is dependent on the instrument operating modes of the observation. The largest coverage at 300 kilometer spacecraft altitude is about 6 km crosstrack and 37 kilometers downtrack. During the Primary Science Phase of MRO operations, HiRISE is expected to image about 1% of the surface of Mars. Observations are carefully selected to optimize science return." END_OBJECT = DATA_SET_INFORMATION OBJECT = DATA_SET_TARGET TARGET_NAME = "MARS" END_OBJECT = DATA_SET_TARGET OBJECT = DATA_SET_HOST INSTRUMENT_HOST_ID = "MRO" INSTRUMENT_ID = "HIRISE" END_OBJECT = DATA_SET_HOST OBJECT = DATA_SET_MISSION MISSION_NAME = "MARS RECONNAISSANCE ORBITER" END_OBJECT = DATA_SET_MISSION OBJECT = DATA_SET_REFERENCE_INFORMATION REFERENCE_KEY_ID = "ELIASON2006A" END_OBJECT = DATA_SET_REFERENCE_INFORMATION OBJECT = DATA_SET_REFERENCE_INFORMATION REFERENCE_KEY_ID = "ELIASON2006B" END_OBJECT = DATA_SET_REFERENCE_INFORMATION END_OBJECT = DATA_SET END