Hello everyone,
I thought I would share this dataset I made recently, perhaps some of you will be interested:
https://www.kaggle.com/romainpessia/artificial-lunar-rocky-landscape-datasetBasically, this is a 10,000 image dataset of somewhat realistic lunar landscapes. As I am an engineer and not an artist, the realism leaves to be desired (both for the ground and for rocks), but the purpose of the dataset is that I also provide the "ground truth" equivalent to the Terragen renders (see picture attached). In other terms, this allows anyone to perform machine learning tasks on "realistic" lunar environments in order to detect rocks.
I hope to see more people use Terragen for scientific purposes in the future, because there is a lot of potential for creating more datasets that don't have real-world equivalents!