Miranda - depth estimation from a single image / test - false color
Visual interpretation of relief on Miranda based on depth estimation from a single image
Zoedepth was not trained for DTM - the 3D image may contain errors, since it's not based on a DTM but on an AI model that hasn't been trained on planets.
Process on 2d image : false color, not RGB
Enlargement, enhancement and colorisation
Crop of a black and white image - cleaned version available on Seti PDS
Surface of Miranda taken by Voyager 2 / NASA - january 24, 1986
Process on 3d image :
Not based on a DTM, but a visual interpretation of the surface
Thank you ZoeDepth: Zero-shot Transfer by Combining Relative and Metric Depth : see it on arxiv.org/abs/2302.12288
Science Credit of image taken by Voyager 2 : NASA/JPL
Choice of processing method (2D/3D) and process execution : Thomas Thomopoulos
Credit for ZoeDepth: Shariq Farooq Bhat, Reiner Birkl, Diana Wofk, Peter Wonka, Matthias Müller
Citation of NASA photojournal about the original image in black and white :
"This image of Miranda, obtained by Voyager 2 on approach, shows an unusual 'chevron' figure and regions of distinctly differing terrain on the Uranian moon. Voyager was 42,000 kilometers (26,000 miles) away when its narrow-angle camera acquired this clear-filter view. Grooved areas baring light and dark bands, distinct from other areas of mottled terrain, are visible at this resolution of about 600 meters (2,000 feet). The bright V-shaped feature in the grooved areas is the 'chevron' observed in earlier, lower-resolution images. Cutting across the bands are sinuous scarps, probably faults. Superimposed on both types of terrain are many bowl-shaped impact craters less than 5 km (3 mi) wide. The entire picture spans an area about 220 km (140 mi) across. The Voyager project is managed for NASA by the Jet Propulsion Laboratory".
Link NASA photojournal : photojournal.jpl.nasa.gov/catalog/PIA00038
Miranda - depth estimation from a single image / test - false color
Visual interpretation of relief on Miranda based on depth estimation from a single image
Zoedepth was not trained for DTM - the 3D image may contain errors, since it's not based on a DTM but on an AI model that hasn't been trained on planets.
Process on 2d image : false color, not RGB
Enlargement, enhancement and colorisation
Crop of a black and white image - cleaned version available on Seti PDS
Surface of Miranda taken by Voyager 2 / NASA - january 24, 1986
Process on 3d image :
Not based on a DTM, but a visual interpretation of the surface
Thank you ZoeDepth: Zero-shot Transfer by Combining Relative and Metric Depth : see it on arxiv.org/abs/2302.12288
Science Credit of image taken by Voyager 2 : NASA/JPL
Choice of processing method (2D/3D) and process execution : Thomas Thomopoulos
Credit for ZoeDepth: Shariq Farooq Bhat, Reiner Birkl, Diana Wofk, Peter Wonka, Matthias Müller
Citation of NASA photojournal about the original image in black and white :
"This image of Miranda, obtained by Voyager 2 on approach, shows an unusual 'chevron' figure and regions of distinctly differing terrain on the Uranian moon. Voyager was 42,000 kilometers (26,000 miles) away when its narrow-angle camera acquired this clear-filter view. Grooved areas baring light and dark bands, distinct from other areas of mottled terrain, are visible at this resolution of about 600 meters (2,000 feet). The bright V-shaped feature in the grooved areas is the 'chevron' observed in earlier, lower-resolution images. Cutting across the bands are sinuous scarps, probably faults. Superimposed on both types of terrain are many bowl-shaped impact craters less than 5 km (3 mi) wide. The entire picture spans an area about 220 km (140 mi) across. The Voyager project is managed for NASA by the Jet Propulsion Laboratory".
Link NASA photojournal : photojournal.jpl.nasa.gov/catalog/PIA00038