Skip to content

Conditional Generative Adversarial Network deployed to predict how satellite images will look like in 10 years based on the changes they have undergone in the last 10 years.

Notifications You must be signed in to change notification settings

hibajamal/Futuristic_LANDSTAT_View

Repository files navigation

LANDSTAT data future predictions

Deep Learning Model with joined Computer Vision algorithms deployed for layer enhancement and image preprocessing.

Using the glorious Google Earth Engine API to access satellite images, we save these GeoTIFFS in order to predict what a certain piece of land will look like in a 10 years. The output will be an image showing the model's prediction.

The dataset and trained model can be found here. Once you access the folder:

  • The generator model is found in the models/ folder
  • Dataset is in three folders, three images are available for each place, for 2000, 2010, and 2020.

About

Conditional Generative Adversarial Network deployed to predict how satellite images will look like in 10 years based on the changes they have undergone in the last 10 years.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published