Deep Space Gateway Science Workshop 2018 (LPI Contrib. No. 2063) 3174.pdf

GLOBAL LUNAR TOPOGRAPHY FROM THE DEEP SPACE GATEWAY FOR SCIENCE AND EXPLORATION. B. Archinal, L. Gaddis, R. Kirk, K. Edmundson, T. Stone, and D. Portree, Astrogeology Science Center, U.S. Geological Survey, Flagstaff, AZ 86001 ([email protected])

Introduction: When we decided to add the scientific However, the later data are from panoramic cameras, instrument module bay to Apollo missions 15 through and as such have some jitter-like distortions similar to 17, photography from orbit was high on our list of sci- those of line scanner cameras. The JAXA Kaguya mis- entific objectives. The Apollo metric camera system was sion also considered collection of such data a priority, flown to acquire photographic data with high accuracy with its TC line scanner cameras obtaining global 10 m to aid the effort of Moon mapping, both for operational stereo imaging [http://www.kaguya.jaxa.jp/en/equip- reasons and for future study and research. - Rocco A. ment/tc_e.htm]. The TC data provide the best near Petrone, Apollo Program Director [1]. global stereo and therefore topographic coverage, and We propose that the Deep Space Gateway (DSG) provides post spacing of ~25-30 m, but are still not ge- carry a new instrument that will finally realize the goal odetically controlled and only approximately registered of Petrone and other NASA planners – and indeed many to the LOLA topography and reference frame [5]. The other investigators and space agencies – to generate a ISRO Chandrayaan-1 mission TMC [https://ti- uniform high resolution global topographic model of the nyurl.com/TMC-ISRO] operated in a similar mode us- Moon. The Apollo Metric and Panoramic cameras in the ing 5 m GSD line scanner cameras, and provide cover- end acquired sufficient data for such coverage over age over some tens of percent of the Moon. It is not clear about 20-25% of the Moon, although only recently have how much of those data have been controlled or pro- the topographic products from those data begun to be cessed into topographic models. The LOLA topo- completed (e.g. [2]). However, even with many new graphic model provides global topographic data over the spectacular lunar datasets, it is still not possible to gen- entire lunar surface, including shadowed areas, and does erate such a model over most of the lunar surface. so in a global reference frame that is probably accurate Such an instrument on the DSG could collect global to the 15-25 m level. The LOLA gridded topography stereo imaging from low lunar orbit for the generation model is derived from along track sampling of ~10-12 of a global uniform high resolution (5 m/pixel imaging, m, but in low to mid latitudes the spacing between tracks 15 m post-spacing) lunar topographic model. Such a can still be several hundred meters. The alignment of the model would fulfill long stated science and exploration LOLA track data has been improved via cross over so- needs. Primarily, it would address a critical need to or- lutions and registration of new data onto existing data thorectify all imaging datasets to a level of high resolu- [6]; however, few meter elevation discrepancies often tion. It would fulfill science needs to properly process persist between adjacent tracks, making the topographic color, multi-, and hyper-spectral data [3], for the best models problematic (e.g. for orthoprojection of high res- possible understanding of lunar surface composition olution images, for slope determination, and illumina- and processes, facilitating global geologic mapping and tion reconstruction) at high resolution in the cross track other lunar surface structural and composition investi- direction. The LROC Narrow Angle Camera (NAC, part gations. It would also support landing site selection and of the LRO Camera (LROC) instrument) has been used surface operations of any science and exploration mis- to obtain repeat stereo coverage and for construction of sions to any illuminated area of the Moon. It would fur- many topographic DTMs at high resolution, with GSD ther allow for the identification of and location of lunar of ~50-200 cm, and post spacing of 1.5-6 m [7]. How- surface changes from crater formation, landslides, and ever, such coverage only exists for a few percent of the boulder movement. Moon, and even with many more years of LRO opera- Past high-resolution lunar topographic datasets: tion will probably only approach several percent of the The availability of global topographic information at Moon. Some of the images and therefore resulting topo- landing site scales has always been considered critical. graphic models are also affected by jitter. As noted, this was a primary goal of the latter Apollo Operations: The proposed instrument would use missions, where coverage was achieved for 20-25% of fore and aft pointing framing cameras, with a GSD from the Moon. In this zone, the Metric camera provided ~7 a nominal 50 km orbit of 5 m, thus providing topo- m ground sample distance (GSD) photographs which graphic post spacing of 2.5-3 times that (12.5-15 m), have been digitized [http://apollo.sese.asu.edu/] and re- and expected vertical precision of 1.4 m which would cently processed into ~40 m GSD topographic and mo- provide global characterization of illuminated terrain at saic information (e.g. [4]). Similar work to process the landing site and surface operations scales. Each of the Panoramic camera data, with ~1-3 m GSD stereo cover- two identical cameras would consist of 4Kx4K CCD age, is now underway for Apollo 15 data at USGS [2]. imaging sensors (using either an active shutter or frame Deep Space Gateway Science Workshop 2018 (LPI Contrib. No. 2063) 3174.pdf

transfer shuttering) and a ~20 cm telescope. Such an im- used to create a Mercury global topographic model, us- age scale could be used to support the location of crater ing USGS ISIS3 software [11], and techniques available and boulder hazards at that (5 m) scale and where since the first (low resolution) global lunar topographic needed, allow reconstruction of topography via photo- model from Clementine stereo data [12]. clinometry (e.g. [8]) at the same scale. This concept fol- Other options: The expected GSD and topographic lows from historic aerial cameras, a previous proposal model resolution could be improved by scaling up to a for such a lunar camera [9], the successful DLR HRSC larger instrument (telescope), but with a consequent in- camera currently operating at Mars [https://ti- crease in observing time and data volume. An additional nyurl.com/HRSC-DLR], and the lunar TC and TMC higher resolution nadir pointing camera could also be cameras. used to do high-resolution mapping in selected loca- In a significant change from the previously flown tions, e.g. completing or extending the same type of cameras, we argue that framing cameras should be used, work that the LROC cameras are currently doing. Some rather than line scanner cameras to largely eliminate ef- overlap in operations with LRO would be expected. It fects of spacecraft jitter. Jitter, of unknown but possibly would also be useful to have a single or multi-beam lidar large amplitude, seems likely on the DSG, due to the instrument (similar to LOLA) in operation to better con- movement of gyros, the large solar electric propulsion nect the stereo images to the LOLA (and/or new lidar solar panels, antennas, hydrazine thrusters, other instru- instrument) reference frame. ments, and of astronauts when occupied. Eliminating Rigorous combined photogrammetric solutions image distortion due to jitter is extremely important for could also be used to process all available data simulta- terrain relative navigation systems that use image corre- neously to improve the global lunar reference frame lation matching. This is currently a critical issue for down to the resolution of the highest resolution global Mars landing site mapping for Mars 2020, since high dataset. The resulting global topographic model would resolution image coverage there is from line scanner be comparable if not better in resolution and positional cameras (HiRISE and CTX) [Yang Chen and Richard accuracy than those for the entire Earth. Otero, personal comm]. Staring at a given location to Benefits: After over 40 years of planning and in- achieve higher SNR in low light (e.g. polar) conditions complete attempts, such an instrument would finally will also be possible. make available a global topographic model that would Complete stereo mapping of the Moon can be ac- serve as the basis or the foundational topographic prod- complished in two months, so the time required for the uct [13] of all lunar orbital and surface observations, for DSG to be in low lunar orbit would not be extensive. any and all scientific and exploration purposes. Such a Extending the observation time or making a return visit model would serve literally as the basis for such pur- might be desirable to fill in gaps, obtain coverage of po- poses for decades, and would likely only need to be up- lar or other areas under better lighting conditions, or ob- dated due to surface changes from impacts and robotic tain higher resolution data from a lower orbit. It is as- and human operations on the Moon. And not inci- sumed that the two cameras would be rigidly fixed at a dentally, the goal of many lunar missions, dating back 40° stereo angle, but that the entire instrument would be to the end of the Apollo missions, would finally be place on a scan platform, with or without other instru- achieved. ments, to avoid reorienting the entire Gateway to point References: [1] Masursky, et al. (1978) NASA SP- the camera. Camera pointing would nominally be to- 362. [2] Edmundson et al. (2017) LPS LXVIII, #2140. ward the nadir, but off-nadir pointing would be neces- [3] Keszthelyi et al. this conf. [4] Edmundson et al. sary to fill in areas missed for operational reasons, and (2016) ISPRS Congress, doi:10.5194/isprs-archives- for off-nadir target of opportunity images. XLI-B4-375-2016. [5] Barker, et al. (2016) Icarus, 273, The overall instrument could be repaired or up- 346. [6] Zuber et al. (2010) SSR, 150, 63. [7] Henriksen graded in the future when human visits occur. Filters et al. (2017) Icarus, 283, 122. [8] Kirk et al. (2003) could be changed or inserted to obtain global stereo im- ISPRS WG IV/9 Workshop, https://tinyurl.com/Simple- aging in color. Astronauts could also operate the cam- Photocl. [9] Davies et al. (1986) Contr. of a Lunar Ge- eras to image targets of opportunity or areas on the sur- oscience Observer Mission…, R. Phillips, ed., SMU, face (e.g. at the lunar poles, once per orbit) where ro- 61. [10] Moratto et al. (2010) LPS XLI, #2364. [11] botic or human operations are taking place. Becker et al. (2016) LPS XLVII, #2959. [12] Archinal Processing of the stereo images into topographic et al. (2006) USGS Open-File Report 2006-1367. [13] models could be accomplished for the global dataset us- Laura et al. (2017) ISPRS doi:10.3390/ijgi6060181. ing established techniques and software, such as the Ames Stereo Pipeline [10], the stereo matching method