OrbitSense submission fields

PROJECT NAME

OrbitSense

SUMMARY

OrbitSense compares Earth terrain with defined reference windows near Shackleton Crater on the Moon and Jezero Crater on Mars. It processes NASADEM, LOLA and MOLA elevation observations on a common metric grid. Users can explore candidate areas and comparison controls on a map, adjust slope, roughness and relief weights, inspect source measurements, and download results. In this catalogue, Cima Volcanic Field ranks first for Jezero. The lunar comparison returns weak agreement, with the Himalaya control above every candidate. Evaluation compares the weighted method with a relief-only baseline and measures ranking changes across feature weights, window sizes and grid resolutions. Each result explains its measured agreements, differences and possible research use. The project supports initial screening of terrestrial field settings and shows where further observations are needed. Its scores describe terrain agreement within a selected catalogue, with documented limits from spatial resolution and the small evaluation set.

PROJECT DETAILS

Problem and intended use

Terrestrial analog research depends on the property and experiment being studied. A familiar site name or similar-looking image gives limited evidence about the terrain of an exact planetary reference area. OrbitSense makes that comparison explicit and lets the user inspect how measurements and priorities affect the ranking.

Working solution

The website compares 13 fixed Earth windows, comprising ten candidate areas and three labelled comparison controls, with two planetary reference windows. The interface provides an Earth map, target selection, adjustable feature weights, ranked comparisons, source measurements, explanations, and CSV or JSON downloads. Processing runs locally in Python. The website loads saved feature files and recalculates scores in the browser.

Exact reference areas

The lunar reference is a 25.6 km square centred at 89.67 degrees south, 129.78 degrees east near Shackleton. It includes crater, rim and exterior terrain. The Martian reference is a 25.6 km square centred at 18.4447 degrees north, 77.4508 degrees east near Jezero's western floor and the Perseverance landing area. Each square spans minus 12,800 to plus 12,800 metres in a local azimuthal equidistant projection. Saved polygons and coordinate-system definitions document the footprints. These are mixed regional footprints. Geological unit boundaries are outside the measurements used here.

Data and physical method

Earth elevations come from NASADEM HGT v001. All selected Earth windows lie within its 60 degrees north to 56 degrees south coverage. The lunar reference uses NASA's adjusted 80 m LOLA south-polar DEM in place of the global 118 m product. The Mars reference uses the MGS MOLA global DEM with nominal 463 m sampling, read through the USGS public S3 mirror because the original file endpoint returned HTTP 403 during acquisition. Source records explain coverage, native coordinate systems, grid spacing, interpolation, acquisition information, and crop hashes.

The pipeline area-averages elevations to a common 800 m metric grid before deriving descriptors. It uses WGS84 for Earth, a 1,737,400 m lunar sphere, and a 3,396,190 m Martian sphere. The three core descriptors are median gradient slope in degrees, RMS height residual after removing a fitted plane in metres, and relative relief defined as the 95th minus 5th elevation percentile in metres. It excludes missing elevation and contaminated gradients. A window needs at least 95 percent valid elevation cells, adequate valid gradient support, and all three core features to receive a score.

For each feature, similarity is exp(-absolute difference / tolerance). The score is 100 times the weighted mean of feature similarities. Weights are normalized. Tolerances use twice the robust spread of reference measurements across the nominal window and its perturbations, with declared floors of 1 degree for slope, 100 m for roughness, and 250 m for relief. These floors are design scales. The interface shows the actual target tolerances. Users can inspect the score contribution of each descriptor.

Measured findings and evaluation

With equal feature weights, Cima Volcanic Field ranks first for the Jezero reference at 63.78/100, followed by El Pinacate at 52.59 and Atacama Yungay at 46.59. For Shackleton, the Himalaya comparison control ranks first at 15.87. The highest candidate score is Mauna Kea at 1.73. The selected catalogue therefore provides weak agreement with this lunar reference under the declared model. The control remains in the results.

The relief-only baseline returns the same top three sites as the full method for each target. This release has no demonstrated shortlist improvement over that baseline. Across 200 deterministic Dirichlet weight draws, mean top-three overlap with the equal-weight ranking is 1.00 for Shackleton and 0.90 for Jezero. A 19.2 km window changes one member of each top-three list. A 1,600 m grid changes one member of Jezero's top three. The public evaluation includes all window and position variants, scores, rank correlations, and per-site rank acceptability.

The evaluation also compares documented analog reference cases with the labelled controls. Existing NASA descriptions document specific geological, environmental or terrain relationships. Those descriptions do not establish binary ground-truth labels for these exact reference windows. Atacama's documented scientific use and Kilauea's lunar geology context remain separate from their measured terrain ranks. The small selected catalogue supports descriptive checks and sensitivity analysis. It does not establish a general prediction-accuracy rate.

Benefits, impact and creative contribution

The project brings actual planetary and Earth terrain observations into one inspectable comparison. Users can change a scientific priority, see the ranking update, and trace a result to its measurements and source crop. This can support initial selection of field settings for geological reconnaissance or regional terrain-planning research. Educators can also use the controls to explain why a partial analogy can hold for one property and fail for another. The contribution is the working integration of reproducible measurements, visible scoring choices, sensitivity results, and property-specific explanations.

Tools and design choices

Python, NumPy, Rasterio, GDAL, pyproj, Requests and Pillow support acquisition and preprocessing. The interface uses HTML, CSS, JavaScript and Leaflet. OpenStreetMap supplies the Earth basemap. Local preprocessing and saved files keep the public app lightweight. The source includes pinned processing dependencies, calculation and interface tests, reproducible source records, and Azure Static Web Apps deployment configuration. Deployment credentials are kept outside the source.

Limitations and further work

The catalogue is a purposive sample. Its rankings apply to the displayed windows, descriptors, physical scale and tolerance model. NASADEM is a radar-derived surface model and may include vegetation or structures. Its February 2000 observations can differ from current volcanic or built terrain. MOLA contains interpolated areas between laser tracks, and LOLA's effective measurement support varies. This release has no bare-ground eligibility mask. Controls with vegetation or cultivated surfaces remain clearly labelled. The regional analysis does not establish habitability, landing safety, construction suitability, wheel-scale mobility, material identity or field access. Further work would add mapped geological units, finer observations and independently labelled reference cases.

PROJECT DEMONSTRATION

https://www.arminnabaei.com/static/orbitsense/demo.html

PROJECT

https://www.arminnabaei.com/orbitsense.html

NASA DATA

NASADEM Merged DEM Global 1 arc second V001
https://doi.org/10.5067/MEaSUREs/NASADEM/NASADEM_HGT.001
NASA JPL, 2020. NASA LP DAAC. Earth elevation observations used for all thirteen Earth windows.

LRO LOLA adjusted south-polar DEM, 80 m
https://pgda.gsfc.nasa.gov/products/90
https://pgda.gsfc.nasa.gov/data/LOLA_20mpp/LDEM_80S_80MPP_ADJ.TIF
Barker et al., 2023, The Planetary Science Journal 4, 183. NASA GSFC Planetary Geodesy Data Archive and LRO LOLA.
https://doi.org/10.3847/PSJ/acf3e1

Mars MGS MOLA DEM 463m
https://astrogeology.usgs.gov/search/map/mars_mgs_mola_dem_463m
https://asc-mars.s3.us-west-2.amazonaws.com/basemaps/Mars_MGS_MOLA_DEM_mosaic_global_463m.tif
NASA Mars Global Surveyor MOLA team and USGS Astrogeology Science Center. Mars elevation observations used for the Jezero reference.

NASA Planetary Analog Explorer
https://science.nasa.gov/solar-system/analog-explorer/
NASA Science. Contextual evidence for specific terrestrial analog research relationships. No gallery photographs are redistributed.

SPACE AGENCY PARTNER & OTHER DATA

Microsoft Planetary Computer
https://planetarycomputer.microsoft.com/dataset/nasadem
Public data catalogue and access route for NASADEM.

OpenStreetMap
https://www.openstreetmap.org/copyright
Earth basemap data copyright OpenStreetMap contributors, available under the Open Database License. Attribution is displayed on the map.

Leaflet 1.9.4
https://leafletjs.com/
https://github.com/Leaflet/Leaflet/blob/v1.9.4/LICENSE
Interactive map library. Its BSD 2-Clause license is retained with the distributed library.

Python
https://www.python.org/

NumPy
https://numpy.org/

Rasterio and GDAL
https://rasterio.readthedocs.io/
https://gdal.org/

pyproj and PROJ
https://pyproj4.github.io/pyproj/
https://proj.org/

Requests
https://requests.readthedocs.io/

Pillow
https://python-pillow.org/

FFmpeg
https://ffmpeg.org/
Encoding the short demonstration from actual application captures.

Microsoft Azure Static Web Apps
https://learn.microsoft.com/azure/static-web-apps/
Optional deployment configuration and setup instructions included in the source project.

Project source and full credits
https://www.arminnabaei.com/static/orbitsense/source.zip
https://www.arminnabaei.com/static/orbitsense/credits.txt

Flask and Jinja
Role: portfolio page integration and static asset delivery.
https://flask.palletsprojects.com/
https://jinja.palletsprojects.com/
