14 research outputs found
Impacts drive lunar rockfalls over billions of years
No full textPast exploration missions have revealed that the lunar topography is eroded through mass wasting processes such as rockfalls and other types of landslides, similar to Earth. We have analyzed an archive of more than 2 million high-resolution images using an AI and big data-driven approach and created the first global map of 136.610 lunar rockfall events. Using this map, we show that mass wasting is primarily driven by impacts and impact-induced fracture networks. We further identify a large number of currently unknown rockfall clusters, potentially revealing regions of recent seismic activity. Our observations show that the oldest, pre-Nectarian topography still hosts rockfalls, indicating that its erosion has been active throughout the late Copernican age and likely continues today. Our findings have important implications for the estimation of the Moon’s erosional state and other airless bodies as well as for the understanding of the topographic evolution of planetary surfaces in general
Automated Discovery of Anomalous Features in Ultralarge Planetary Remote-Sensing Datasets Using Variational Autoencoders
No full textThe NASA Lunar Reconnaissance Orbiter (LRO) has returned petabytes of lunar high spatial resolution surface imagery over the past decade, impractical for humans to fully review manually. Here, we develop an automated method using a deep generative visual model that rapidly retrieves scientifically interesting examples of LRO surface imagery representing the first planetary image anomaly detector. We give quantitative experimental evidence that our method preferentially retrieves anomalous samples such as notable geological features and known human landing and spacecraft crash sites. Our method addresses a major capability gap in planetary science and presents a novel way to unlock insights hidden in ever-increasing remote-sensing data archives, with numerous applications to other science domains
Quantitative Assessment of Digital Image Correlation Methods to Detect and Monitor Surface Displacements of Large Slope Instabilities
No full textWe evaluate the capability of three different digital image correlation (DIC) algorithms to measure long-term surface displacement caused by a large slope instability in the Swiss Alps. DIC was applied to high-resolution optical imagery taken by airborne sensors, and the accuracy of the displacements assessed against global navigation satellite system measurements. A dynamic radiometric correction of the input images prior to DIC application was shown to enhance both the correlation success and accuracy. Moreover, a newly developed spatial filter considering the displacement direction and magnitude proved to be an effective tool to enhance DIC performance and accuracy. Our results show that all algorithms are capable of quantifying slope instability displacements, with average errors ranging from 8 to 12% of the observed maximum displacement, depending on the DIC processing parameters, and the pre- and postprocessing of the in- and output. Among the tested approaches, the results based on a fast Fourier transform correlation approach provide a considerably better spatial coverage of the displacement field of the slope instability. The findings of this study are relevant for slope instability detection and monitoring via DIC, especially in the context of an ever-increasing availability of high-resolution air- and spaceborne imagery
Automated astronaut traverses with minimum metabolic workload: Accessing permanently shadowed regions near the lunar South pole
No full textAltres ajuts: acords transformatius de la UABThe Artemis exploration zone is a topographically complex impact-cratered terrain. Steep undulating slopes pose a challenge for walking extravehicular activities (EVAs) anticipated for the Artemis III and subsequent missions. Using 5 m/pixel Lunar Orbiter Laser Altimeter (LOLA) measurements of the surface, an automated Python pipeline was developed to calculate traverse paths that minimize metabolic workload. The tool combines a Monte Carlo method with a minimum-cost path algorithm that assesses cumulative slope over distances between a lander and stations, as well as between stations. To illustrate the functionality of the tool, optimized paths to permanently shadowed regions (PSRs) are calculated around potential landing sites 001, nearby location 001(6), and 004, all within the Artemis III 'Connecting Ridge' candidate landing region. We identified 521 PSRs and computed (1) traverse paths to accessible PSRs within 2 km of the landing sites, and (2) optimized descents from host crater rims into each PSR. Slopes are limited to 15° and previously identified boulders are avoided. Surface temperature, astronaut body illumination, regolith bearing capacity, and astronaut-to-lander direct view are simultaneously evaluated. Travel times are estimated using Apollo 12 and 14 walking EVA data. A total of 20 and 19 PSRs are accessible from sites 001 and 001(6), respectively, four of which maintain slopes <10°. Site 004 provides access to 11 PSRs, albeit with higher EVA workloads. From the crater rims, 94 % of PSRs can be accessed. All round-trip traverses from potential landing sites can be performed in under 2 h with a constant walk. Traverses and descents to PSRs are compiled in an atlas to support Artemis mission planning
Systematic Quantification and Assessment of Digital Image Correlation Performance for Landslide Monitoring
No full textAccurate and reliable analyses of high-alpine landslide displacement magnitudes and rates are key requirements for current and future alpine early warnings. It has been proved that high spatiotemporal-resolution remote sensing data combined with digital image correlation (DIC) algorithms can accurately monitor ground displacements. DIC algorithms still rely on significant amounts of expert input; there is neither a general mathematical description of type and spatiotemporal resolution of input data nor DIC parameters required for successful landslide detection, accurate characterisation of displacement magnitude and rate, and overall error estimation. This work provides generic formulas estimating appropriate DIC input parameters, drastically reducing the time required for manual input parameter optimisation. We employed the open-source code DIC-FFT using optical remote sensing data acquired between 2014 and 2020 for two landslides in Switzerland to qualitatively and quantitatively show which spatial resolution is required to recognise slope displacements, from satellite images to aerial orthophotos, and how the spatial resolution affects the accuracy of the calculated displacement magnitude and rate. We verified our results by manually tracing geomorphic markers in orthophotos. Here, we show a first generic approach for designing and optimising future remote sensing-based landslide monitoring campaigns to support time-critical applications like early warning systems
Superresolution of Lunar Satellite Images for Enhanced Robotic Traverse Planning: Maximizing the Value of Existing Data Products for Space Robotics
No full text100112Lunar exploration missions require detailed and accurate planning to ensure their safety. Remote sensing data, such as optical satellite imagery acquired by lunar orbiters, are key for the identification of future landing and mission sites. Here robot- and astronaut-scale obstacles are the most relevant to resolve; however, the spatial resolution of the available image data is often insufficient, particularly in the poorly illuminated polar regions of the moon, leading to uncertainty. This work shows how a novel single-image superresolution (SISR) application, the Adversarial Network for Uncertainty-Based Image SR (ANUBIS), can enhance lunar surface imagery by improving the resolution by a factor of two, outperforming other approaches and benchmarks. The enhanced images improve the reliability and detail of lunar traverse planning and topographic reconstruction, while providing an estimate of the uncertainty associated with the enhancement process, vital to ensure mission planning integrity. This work demonstrates how machine-learning-driven processing can enhance existing data products to maximize their value for science and the exploration of the moon and other celestial bodies.31
Deep Learning-driven Detection and Mapping of Rockfalls on Mars
No full textThe analysis of rockfall distribution and magnitude is a useful tool to study the past and current endogenic and exogenic activity of Mars. At the same time, tracks left by rockfalls provide insights into the mechanical properties of the martian surface. While a wealth of high-resolution space-borne image data are available, manual mapping of displaced boulders with tracks is inefficient and slow, resulting in 1) a small total number of mapped features, 2) inadequate statistics, and 3) a sub-optimal utilization of the available big data. This study implements a deep learning-driven approach to automatically detect and map martian boulders with tracks in High Resolution Imaging Science Experiment (HiRISE) imagery. Six off-the-shelf neural networks have been trained either on martian or lunar rockfall data, or a combination of both, and are able to achieve a maximum recall of up to 0.78 and a maximum precision of up to 1.0, with a mean average precision of 0.71. The fusion of training data from different planets and sensors results in an increased detection precision, highlighting the value of domain generalization and multi-domain learning. Average processing time per HiRISE image is ~45 s using an NVIDIA Titan Xp, which is more than one order of magnitude faster than a human operator. The developed deep learning-driven infrastructure can be deployed to map martian rockfalls on a global scale and within a realistic timeframe
A Global Study of Dust Devils on Mars using Neural Networks and Zooniverse
No full textInternational audienc
