Unsupervised Distribution Learning for Lunar Surface Anomaly Detection

Abstract

In this work we show that modern data-driven machine learning techniques can be successfully applied on lunar surface remote sensing data to learn, in an unsupervised way, sufficiently good representations of the data distribution to enable lunar technosignature and anomaly detection. In particular we train an unsupervised distribution learning model to find the landing module of the Apollo 15 landing site in a testing dataset, with no dataset specific model or hyperparameter tuning. Sufficiently good unsupervised data density estimation techniques have the potential to enable a dazzling number of useful downstream tasks, including locating lunar resources for future space flight and colonization, finding new impact craters or lunar surface reshaping, and deciding the importance of unlabeled samples to send back from power-and bandwidth-constrained missions. We show in this work that such unsupervised learning can be successfully done in the lunar remote sensing and space sciences contexts.

Technosignatures: Frameworks for Their Assessment

“…However, another example of technosignatures merits consideration, to wit, the possibility of discovering physical artifacts of extraterrestrial technology within the solar system Bradbury et al 2011;Haqq-Misra & Kopparapu 2012;Lacki 2019;Shostak 2020). This notion was broached by Bracewell (1960) in a well-known publication, after which a handful of searches for artifacts have been conducted in geosynchronous Earth orbits (Villarroel et al 2022a(Villarroel et al , 2022b, at the Earth-Moon Lagrange points (Freitas & Valdes 1980;Valdes & Freitas 1983;Freitas & Valdes 1985), and on the lunar surface (Arkhipov 1995;Lesnikowski et al 2020). Artifact searches could potentially be initiated for objects in proximity to Earth (Steel 1995;Arkhipov 1996;Benford 2021;Loeb & Laukien 2022); co-orbital objects (Benford 2019); or objects located on the surfaces of planets and moons (Carlotto & Stein 1990;Arkhipov et al 1996;Davies & Wagner 2013).…”

Review for Examining the Oxidation Process of the Moon Using Generative Adversarial Networks: Focusing on Landscape of Moon

“…Lesnikowski et al experimented with detecting the landing sites of Apollo 15 and 17 by using a variational autoencoder to show that lunar surface anomaly detection can be performed unsupervised without sufficient representations. In addition, they mentioned that unsupervised data density estimations can be expanded to various tasks, including locating lunar resources [22].…”

“…Lee Honnhee [14] Review Papers (Usually, crater detection) -Jia et al [15] Lunar surface detection Self-calibrated convolution Silburt et al [16] Lunar surface detection CNNs (based U-Net) Yutong Jia et al [17] Lunar surface detection CNNs (based U-Net) Ali-Dib et al [18] Lunar surface detection CNNs (based Mask R-CNN) Shen et al [19] Lunar surface detection High-Resolution-Moon-Net Wilhelm et al [20] Unsupervised learning CNNs (based VGG16) Roy et al [21] Unsupervised learning CNNs (based U-Net) Lesnikowski et al [22] Unsupervised learning CNNs (based VAE) Xia et al [23] Abundance map of oxide and magnesium DNN…”

SETI in 2020

“…Brzycki et al ( 2020) described using convolutional neural networks to identify and characterize narrow band signals in noisy data with an eye towards radio SETI. Lesnikowski et al (2020) described an unsupervised neural network to analyze imagery of the moon to look for artifacts and other anomalies, and shows they can recover the obviously artificial Apollo 15 landing site with it.…”