Martian Earth · 火星照耀地球

When Mars shines on Earth through AI’s imagination...

With most of its atmosphere lost and its crust oxidized to various shades of red, are these views of Earth a representation of its pristine state or a not-so-favorable future that we'll have to face? Or, are the night lights upon the red terrain a blueprint for our Mars development? In Martian Earth, I present artistic renderings of Earth with visual reference to the "red planet"—Mars. The creative process is the reverse of Terra Mars. I have trained an artificial neural network with topographical data and satellite imagery of Mars so that it can learn the relation between them. The trained model was then applied to topographical data of Earth to generate images that resemble satellite imagery of Mars. With recognizable topographical features of Earth and color patterns emblematic of Mars, I hope this AI-powered mental connection can demonstrate the artistic potential of the technology and provoke people's contemplation of our future on an interplanetary level.


Notes

  • Generation: A Martian Earth generation refers to the results of applying the trained model to the full dataset of Earth topography, and stitching the output into a global color map of Earth. The process usually but not always includes training a new model. The generations are numbered in chronological order.

Data credits

  • "Blue Marble: Next Generation," (NASA, 2005).

  • "Mars HRSC MOLA Blended DEM Global 200m v2," (U.S. Geological Survey, 2017).

  • "Viking Global Color Mosaic (MDIM1 controlled) 925m," (U.S. Geological Survey, 2001).

Neural network model credits

Early iterations of the Terra Mars series were based on the following neural network models:

  • P. Isola, et al. "Image-To-Image Translation With Conditional Adversarial Networks," Proceedings of IEEE Conference on Computer Vision and Pattern Recognition 2017, 5967–5976.

  • T. C. Wang, et al. "High-Resolution Image Synthesis and Semantic Manipulation with Conditional GANs," Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition 2018, 8798–8807.

Asset credits

  • Night lights texture: "Black Marble 2016," (NASA, 2017).

  • Stars texture: "Deep Star Maps," (NASA, 2012).


Terra Mars Series

Martian Earth belongs to the Terra Mars series. Projects in the series include:

Terra Mars

2019

Martian Earth

2019