Generation of Simulated "Ultra-High Resolution" HiRISE Imagery Using Enhanced Super-Resolution Generative Adversarial Network Modeling
Abstract
An Enhanced Super-Resolution Genarative Adversarial Network (ESRGAN) model is trained to build ultra-high spatial resolution HiRISE imagery. The images are higly convincing with much sharper outlines and no artifact that could be noticed.
- Publication:
-
52nd Lunar and Planetary Science Conference
- Pub Date:
- March 2021
- Bibcode:
- 2021LPI....52.1935C