Small Craters Research Articles

Heterogeneity across Neotropical aquatic environments affects prokaryotic and eukaryotic biodiversity based on environmental DNA

AbstractCharacterizing biological communities and knowledge on the distribution of biodiversity allows the assessment of ecological quality. This provides valuable information for conservation biology and monitoring purposes. While obtaining such data has been challenging in the past, environmental DNA (eDNA) sampling represents a promising tool to describe biodiversity on a broad taxonomic scale. In this study, we provide the first broad‐scale biodiversity assessment for ten Neotropical water bodies in Nicaragua (a major river, two great lakes, and seven relatively young and small crater lakes) using eDNA sampling to determine how abiotic factors structure the distribution of prokaryotic and eukaryotic biodiversity across these environments. Further, we explored to what extent levels of biodiversity are associated across different taxonomic groups and environments. We found that prokaryotic and eukaryotic α‐diversity was consistently higher in the great lakes (i.e., Lakes Nicaragua and Managua) as well as in Río San Juan compared with the young and small crater lakes. Differences of prokaryotic and eukaryotic communities (β‐diversity) were significantly correlated, indicating that biological communities are similarly structured across environments. Accordingly, differences in salinity were correlated with prokaryotic and eukaryotic communities, whereas differences in dissolved oxygen were only correlated with prokaryotic communities (β‐diversity). Yet, salinity and dissolved oxygen only affected prokaryotic α‐diversity, suggesting different effects of these two abiotic factors on biodiversity within aquatic environments. Moreover, α‐diversity of different phyla was positively correlated, although more strongly in prokaryotes, showing that biodiversity patterns are congruent across a broad range of lineages, particularly in prokaryotes.

The Geological History of the Chang’e-5 Sample Return Region

Chang’e-5 (CE-5), China’s first sample-return mission, has successfully landed in Oceanus Procellarum near Mons Rümker. It is important to have a detailed study of the geological evolution of the CE-5 sample return region. This work aims to study the geological background, topography, geomorphology, major chemical composition, mineralogy, and chronology of the landing site region. First, we used the map of topography obtained by the Kaguya TC merged Digital Terrain Model (DTM) to analyze the topographic characteristics. Then, we used the Kaguya Multiband Imager (MI) reflectance data to derive FeO and TiO2 abundance and the hyperspectral data of the Moon Mineralogy Mapper (M3) onboard the Chandrayaan-1 spacecraft to study the mineralogy of the landing site region. Later, we defined and dated the geological units of the landing area using the crater size–frequency distribution (CSFD) method. Finally, we conducted a detailed analysis of the volcanism and tectonism that occurred in the CE-5 landing area. The study region has experienced multi-stage magmatic activities (~3.36 Ga to ~1.22 Ga) and formed multiple mare units with different chemical and mineral compositions. The relationship between the wrinkle ridges cut by small impact craters suggests that the U7/Em5 has experienced Copernican aged tectonism recently ~320 Ma. The U7/Em5 unit where the Chang’e-5 sample return mission landed is dominantly composed of mature pyroxene and the basalts are mainly high-iron and mid-titanium basalts. Additionally, the analysis of pure basalt in the U7/Em5 suggests that the samples returned by the CE-5 mission may contain the ejecta and ray materials of young craters, including sharp B, Harding, Copernicus, and Aristarchus.

Effect of Topographic Degradation on Small Lunar Craters: Implications for Regolith Thickness Estimation

AbstractSmall crater morphology method has been used extensively in lunar regolith thickness estimation. However, topographic degradation can change crater morphology and thus bias regolith thickness estimation. In this study, we first developed a shape model for small fresh craters with normal, central mound, flat‐bottomed, and concentric geometry. We then simulated their degradation processes by using a topographic diffusion model. Simulation results show that as a small crater degrades, its morphology changes from concentric/central mound to flat‐bottomed, from flat‐bottomed to normal, and from normal to invisible, depending on its initial morphology. Upon the time a crater becomes invisible, its diameter can be enlarged by a factor of ∼70%. We proposed a revised small crater morphology method and applied it to the Apollo 11 and 14 landing sites. Our revised method permits a more accurate estimate of regolith thickness, and our results are helpful in understanding the evolution of the Moon's surface.

High-Spatial Resolution Atmospheric Pressure Mass Spectrometry Imaging Using Fiber Probe Laser Ablation-Dielectric Barrier Discharge Ionization.

Atmospheric pressure mass spectrometry imaging (AP-MSI) is a powerful tool in many fields; however, there are still some difficulties to achieve high spatial resolution for AP-MSI, one of them being the need for a small ablation crater. Here, a fiber probe laser ablation (FPLA) system is introduced that uses an etched optical fiber with a sharp tip (o.d. 200 nm) to deliver ablation laser pulses to a sample surface to ablate materials with high spatial resolution. The tip-to-sample distance was adjusted to ∼10 μm using a micro-actuator having a stepping motor with submicron accuracy. The laser-ablated neutrals were post-ionized using a home-built in-line dielectric barrier discharge source, which can be interfaced to any mass spectrometer with an AP interface. Using MSI on a standard sample with a striped pattern and sections of fingernails treated with the drug methyl green zinc chloride salt, a FPLA-DBDI-MSI spatial resolution of ≈5 μm was demonstrated.

Recognition of Sedimentary Rock Occurrences in Satellite and Aerial Images of Other Worlds—Insights from Mars

Sedimentary rocks provide records of past surface and subsurface processes and environments. The first step in the study of the sedimentary rock record of another world is to learn to recognize their occurrences in images from instruments aboard orbiting, flyby, or aerial platforms. For two decades, Mars has been known to have sedimentary rocks; however, planet-wide identification is incomplete. Global coverage at 0.25–6 m/pixel, and observations from the Curiosity rover in Gale crater, expand the ability to recognize Martian sedimentary rocks. No longer limited to cases that are light-toned, lightly cratered, and stratified—or mimic original depositional setting (e.g., lithified deltas)—Martian sedimentary rocks include dark-toned examples, as well as rocks that are erosion-resistant enough to retain small craters as well as do lava flows. Breakdown of conglomerates, breccias, and even some mudstones, can produce a pebbly regolith that imparts a “smooth” appearance in satellite and aerial images. Context is important; sedimentary rocks remain challenging to distinguish from primary igneous rocks in some cases. Detection of ultramafic, mafic, or andesitic compositions do not dictate that a rock is igneous, and clast genesis should be considered separately from the depositional record. Mars likely has much more sedimentary rock than previously recognized.

Millimeter‐ to Decimeter‐Scale Surface Roughness of the Moon at the Chang'e‐4 Exploration Region

AbstractThe surface slope and roughness of the Moon have been investigated extensively over a wide baseline range except millimeter to decimeter scales. In this study, we present for the first time millimeter‐to decimeter‐scale surface slope and roughness of the Moon at China's Chang'e‐4 landing regions (∼20 m across) using the Digital Terrain Model (DTM) with a resolution of 5 mm/pixel. The bidirectional slope at the 7 mm scale can be larger than 40° with a median value of ∼10°. The root‐mean‐square (RMS) height within a window size of 125 mm varies from ∼1 mm to ∼18 mm with a median value of ∼4 mm. Both the bidirectional slope and RMS height show scale‐dependent behaviors and the parameter of scale dependence, the Hurst exponent, is ∼0.6–0.85. We also synthesized the bidirectional slope at baseline from micrometer to kilometer, showing that bidirectional slope decreases from ∼60° at micrometer to ∼1° at a kilometer. At millimeter‐scale, surface roughness is mainly controlled by small impact craters, rocks, and regolith properties. Our roughness results not only bridge the gap in understanding surface roughness from traditional topographic data sets to radar and thermal observations, but also provide valuable information about lunar regolith characteristics, and small‐scale geological processes.

Crater Detection and Recognition Method for Pose Estimation

A crater detection and recognition algorithm is the key to pose estimation based on craters. Due to the changing viewing angle and varying height, the crater is imaged as an ellipse and the scale changes in the landing camera. In this paper, a robust and efficient crater detection and recognition algorithm for fusing the information of sequence images for pose estimation is designed, which can be used in both flying in orbit around and landing phases. Our method consists of two stages: stage 1 for crater detection and stage 2 for crater recognition. In stage 1, a single-stage network with dense anchor points (dense point crater detection network, DPCDN) is conducive to dealing with multi-scale craters, especially small and dense crater scenes. The fast feature-extraction layer (FEL) of the network improves detection speed and reduces network parameters without losing accuracy. We comprehensively evaluate this method and present state-of-art detection performance on a Mars crater dataset. In stage 2, taking the encoded features and intersection over union (IOU) of craters as weights, we solve the weighted bipartite graph matching problem, which is matching craters in the image with the previously identified craters and the pre-established craters database. The former is called “frame-frame match”, or FFM, and the latter is called “frame-database match”, or FDM. Combining the FFM with FDM, the recognition speed is enabled to achieve real-time on the CPU (25 FPS) and the average recognition precision is 98.5%. Finally, the recognition result is used to estimate the pose using the perspective-n-point (PnP) algorithm and results show that the root mean square error (RMSE) of trajectories is less than 10 m and the angle error is less than 1.5 degrees.

Comprehensive Geophysical Study at Wabar Crater, Rub Al‐Khali Desert, Saudi Arabia

AbstractInterest in impact craters on the earth's surface has increased worldwide and is being investigated by using remote sensing, geological, boreholes, geophysical, and laboratory measurements. These measurements are used to build dynamic models to study crater formation. In this work, the near‐crater sediments at the young Wabar crater field in Saudi Arabia have been investigated using magnetic, transient electromagnetic (TEM), seismic, and ground‐penetrating radar (GPR) methods. The main objectives of this research were to (a) explore the possibility of any remnant major pieces of the meteorite, (b) investigate the meteoroid direction, and (c) map the deformational structures associated with the meteorite impact. Our results show five different magnetic anomaly types and three layers in the subsurface. The maximum depth of deformation due to the impact of the meteorite is about 25 m as shown by the seismic travel time tomogram, the quasi‐2D TEM, and the 3D GPR model. TEM survey confirmed the geometrical characteristics of the major crater and located another small crater (known as Philby‐A). The magnetic survey shows no evidence of any remnant major pieces of the meteorite; however, it was used to trace ejecta material containing highly dilute magnetic material. The magnetic carrier is most likely spheres of metal incorporated in the black/green glasses. During the expedition, many small pieces of the meteoroid were found and collected for further geochemical analysis. Based on the geophysical findings, the meteorite direction was found to be from north to south.

Rock Fragments in Shallow Lunar Regolith: Constraints by the Lunar Penetrating Radar Onboard the Chang'E‐4 Mission

AbstractIt is generally believed that the top centimeters of mature lunar regolith should be dominated by fine particles that contain few rocky fragments, but surface fragments are occasionally visible in mature lunar regolith. Based on amplitudes of surface echoes received by the Lunar Penetrating Radar onboard the Yutu‐2 rover of the Chang'E‐4 mission, we extracted regions with abnormally high relative permittivity that is larger than 4, quantifying the abundance of dense fragments within the top ∼4 cm of the mature regolith along the rover path. We carried out high‐resolution numerical simulations for the propagation of electromagnetic waves, showing that the abnormally high relative permittivity is not caused by local topography but due to the existence of dense materials. Using high‐resolution images obtained by the panorama camera onboard the rover, we performed correlated observation for areas with abnormally high relative permittivity, confirming that some of these dense materials are exposed rocky fragments. However, most of the high relative permittivity areas are not correlated with surface fragments, and some are located beneath floors of highly degraded small impact craters. We found that regolith clumps formed due to impact compaction are not dense enough to cause the abnormally large relative permittivity. Alternatively, the detected dense materials may be melt‐bearing impact breccias in the shallow regolith, which were formed by the host craters. The results revealed a highly heterogeneous structure at shallow depths of mature lunar regolith, revealing a new mechanism of sand‐blasting impacts on the destruction of pores in mature lunar regolith.

An Improved Global Catalog of Lunar Impact Craters (≥1 km) With 3D Morphometric Information and Updates on Global Crater Analysis

AbstractImpact craters are common surface features on planetary surfaces. Their distribution offers important clues regarding geological and temporal processes on the Moon. Numerous endeavors have generated global catalogs of lunar craters; however, most of the existing catalogs only contain large craters (larger than several kilometers in diameter), and none of them offer three‐dimensional (3D) morphometric information. In this study, we first present a machine‐learning approach for automatic crater detection from digital elevation models (DEMs). Our crater detection approach can achieve a detection rate of about 85%. We also present an approach for extracting 3D morphometric information of craters based on the topography. These approaches were applied to producing a global crater data set covering the entire lunar surface that includes approximately 1.32 million craters (≥1 km). Verification was performed against previously published catalogs (Head et al., 2010, https://doi.org/10.1126/science.1195050; Robbins, 2019, https://doi.org/10.1029/2018je005592), with about 23% of our craters transferred from the previous catalogs after a rim‐fitting process. The crater catalog includes 3D morphometric data on the craters such as depths. Global analyses of craters based on this improved catalog indicate that the lunar mare and highlands have distinctive crater density differences of small craters (1–5 km). Craters of 2.5–5 km have reached saturation in several local regions in the highlands. Small craters (1–5 km) on the lunar mare are deeper than those on the highlands. The data in our comprehensive crater catalog can support various other lunar scientific studies.

Effect of Topographic Degradation on Small Lunar Craters: Implications for Regolith Thickness Estimation

Effect of Topographic Degradation on Small Lunar Craters: Implications for Regolith Thickness Estimation

Making an Onboard Reference Map From MRO/CTX Imagery for Mars 2020 Lander Vision System

AbstractThe Mars 2020 rover, Perseverance, landed in Jezero crater (18.4663°N, 77.4298°E) on February 18, 2021 to collect samples from Mars that could be returned to Earth by a future Mars Sample Return campaign. While providing a rich sampling opportunity, Jezero also contains numerous landing hazards including scarps, canyons, mesas, dune fields, rock fields, and smaller craters. The Mars 2020 onboard inertial navigation system, which is the same as that used by the Mars Science Laboratory (MSL), only provides a very coarse inertially propagated position, which can have error as large as 3.2 km. The Lander Vision System (LVS) was added to Mars 2020 to reduce this position knowledge error to less than 40 m with respect to an on‐board reference map of the landing area. LVS uses a reference map on board to compare with the descent images for lander localization during the terminal stage of Entry, Descent and Landing (EDL). Because the reference map is used directly during EDL, it is critical that it have as little spatial and photometric error as possible. Photometrically, it should resemble as much as possible the real descent images to allow reliable terrain matching. Spatially, it should match as faithfully as possible the real, underlying terrain and contribute minimal error to the final localization solution. On February 18, 2021, the LVS reference map passed its ultimate test. The LVS system executed Terrain Relative Navigation (TRN) flawlessly based on the LVS reference map. In addition to its use for TRN, the local hazard map is also registered to the LVS reference map so that the safe target selection (STS) system can select a safe and reachable landing site. The final error between the site targeted by the STS system and the real landing site is estimated at only 5 m. The exceptional performance of LVS indicates that the reference map met mission requirements with comfortable margin. LVS on Mars 2020 represents the first ever use of a reference map during spacecraft EDL. This breakthrough will have profound implications for future lander missions and the scope of scientific inquiry they are able to address. In this paper, we will describe the necessary precursor steps to building the Jezero Crater LVS map, including the methodology to dejitter Context Imager (CTX) images, improve the CTX sensor model, as well as the process used to validate the LVS map accuracy.

The 2018 phreatic eruption at Mt. Motoshirane of Kusatsu\u2013Shirane volcano, Japan: eruption and intrusion of hydrothermal fluid observed by a borehole tiltmeter network

We estimate the mass and energy budgets for the 2018 phreatic eruption of Mt. Motoshirane on Kusatsu–Shirane volcano, Japan, based on data obtained from a network of eight tiltmeters and weather radar echoes. The tilt records can be explained by a subvertical crack model. Small craters that were formed by previous eruptions are aligned WNW–ESE, which is consistent with the strike of the crack modeled in this study. The direction of maximum compressive stress in this region is horizontal and oriented WNW–ESE, allowing fluid to intrude from depth through a crack with this orientation. Based on the crack model, hypocenter distribution, and MT resistivity structure, we infer that fluid from a hydrothermal reservoir at a depth of 2 km below Kusatsu–Shirane volcano has repeatedly ascended through a pre-existing subvertical crack. The inflation and deflation volumes during the 2018 eruption are estimated to have been 5.1 × 105 and 3.6 × 105 m3, respectively, meaning that 1.5 × 105 m3 of expanded volume formed underground. The total heat associated with the expanded volume is estimated to have been ≥ 1014 J, similar to or exceeding the annual heat released from Yugama Crater Lake of Mt. Shirane and that from the largest eruption during the past 130 year. Although the ejecta mass of the 2018 phreatic eruption was small, the eruption at Mt. Motoshirane was not negligible in terms of the energy budget of Kusatsu–Shirane volcano. A water mass of 0.1–2.0 × 107 kg was discharged as a volcanic cloud, based on weather radar echoes, which is smaller than the mass associated with the deflation. We suggest that underground water acted as a buffer against the sudden intrusion of hydrothermal fluids, absorbing some of the fluid that ascended through the crack.

Geologic History and Crater Morphology of Asteroid (162173) Ryugu

AbstractCrater morphology and surface age of asteroid (162173) Ryugu are characterized using the high‐resolution images obtained by the Hayabusa2 spacecraft. Our observations reveal that the abundant boulders on and under the surface of the rubble‐pile asteroid affect crater morphology. Most of the craters on Ryugu exhibit well‐defined circular depressions, unlike those observed on asteroid Itokawa. The craters are typically outlined by boulders remaining on the rim. Large craters (diameter > 100 m) host abundant and sometimes unproportionally large boulders on their floors. Small craters (<20 m) are characterized by smooth circular floors distinguishable from the boulder‐rich exterior. Such small craters tend to have dark centers of unclear origin. The correlation between crater size and boulder number density suggests that some processes sort the size of boulders in the shallow (<30 m) subsurface. Furthermore, the crater size‐frequency distributions (CSFDs) of different regions on Ryugu record multiple geologic events, revealing the diverse geologic history on this 1‐km asteroid. Our crater‐counting analyses indicate that the equatorial ridge is the oldest structure of Ryugu and was formed 23–30 Myr ago. Then, Ryugu was partially resurfaced, possibly by the impact that formed the Urashima crater 5–12 Myr ago. Subsequently, a large‐scale resurfacing event formed the western bulge and the fossae 2–9 Myr ago. Following this process, the spin of Ryugu slowed down plausibly due to the YORP effect. The transition of isochrons in a CSFD suggests that Ryugu was decoupled from the main belt and transferred to a near‐Earth orbit 0.2–7 Myr ago.

Effect of Topographic Degradation on Small Lunar Craters: Implications for Regolith Thickness Estimation

Effect of Topographic Degradation on Small Lunar Craters: Implications for Regolith Thickness Estimation

Semi-Supervised Deep Learning for Lunar Crater Detection Using CE-2 DOM

Lunar craters are very important for estimating the geological age of the Moon, studying the evolution of the Moon, and for landing site selection. Due to a lack of labeled samples, processing times due to high-resolution imagery, the small number of suitable detection models, and the influence of solar illumination, Crater Detection Algorithms (CDAs) based on Digital Orthophoto Maps (DOMs) have not yet been well-developed. In this paper, a large number of training data are labeled manually in the Highland and Maria regions, using the Chang’E-2 (CE-2) DOM; however, the labeled data cannot cover all kinds of crater types. To solve the problem of small crater detection, a new crater detection model (Crater R-CNN) is proposed, which can effectively extract the spatial and semantic information of craters from DOM data. As incomplete labeled samples are not conducive for model training, the Two-Teachers Self-training with Noise (TTSN) method is used to train the Crater R-CNN model, thus constructing a new model—called Crater R-CNN with TTSN—which can achieve state-of-the-art performance. To evaluate the accuracy of the model, three other detection models (Mask R-CNN, no-Mask R-CNN, and Crater R-CNN) based on semi-supervised deep learning were used to detect craters in the Highland and Maria regions. The results indicate that Crater R-CNN with TTSN achieved the highest precision (of 91.4% and 88.5%, respectively) in the Highland and Maria regions, even obtaining the highest recall and F1 score. Compared with Mask R-CNN, no-Mask R-CNN, and Crater R-CNN, Crater R-CNN with TTSN had strong robustness and better generalization ability for crater detection within 1 km in different terrains, making it possible to detect small craters with high accuracy when using DOM data.

Experimental research on single-pulse discharge crater morphology in SEAM

ABSTRACT The analysis of craters in electrical discharge machining (EDM) can help improve the understanding of the basic theory and predict the process indexes; thus, the research on craters has been a popular topic in this field. However, a novel low-voltage, high-current EDM technique based on a constant-voltage power supply, namely short electric arc machining (SEAM), is almost completely unexplored. To fill this gap and reveal the basic theory of SEAM, in this work, the influences of the voltage and pulse duration under different polarities on the crater diameter, the width, depth, and thickness of the recast layer, and the bulge height are investigated by single-factor experiments using single-pulse discharge, and the differences between the surfaces and cross-sections of craters are further analyzed. The experimental results demonstrate that as the voltage and pulse duration increase, the discharge channel expands, the energy density of the discharge channel increases, and the crater diameter, the width and depth of the recast layer, and the bulge height all increase. It is also found that negative-polarity machining tends to form large and shallow craters with lower surface roughness, while positive-polarity machining tends to form small and deep craters with higher machining efficiency.

Regolith Properties in the Chang'E‐5 Landing Region of the Moon: Results From Multi‐Source Remote Sensing Observations

AbstractIn December 2020, China's Chang'E‐5 (CE‐5) spacecraft successfully collected 1.731 kg samples from Oceanus Procellarum on the Moon's nearside, becoming the first lunar sample return mission since Luna 24 in 1976. These samples, collected from one of the youngest mare basalt units within the Procellarum KREEP Terrain, are critical to deciphering several fundamental questions of the Moon. In order to provide geologic context for interpreting the returned samples, we conducted a comprehensive analysis of regolith properties in the landing region using optical, topographic, multispectral, thermal, and radar observations. The CE‐5 landing site is within a flat, young mare basalt unit with intermediate titanium. A ballistic sedimentation model suggests that ∼98% of the surface regolith are local materials. The dielectric permittivity of the regolith is estimated to be 2.96 + i0.03 based on its composition and porosity. Morphologies of small fresh craters identified in high‐resolution optical images show that regolith thickness varies from ∼1.5 to ∼8 m with a median value of ∼5 m, implying a large mean regolith growth rate over the landing region. A comparison between radar image and Lunar Reconnaissance Orbiter Diviner surface rock abundance (RA) map indicates that subsurface rocks play a significant role in producing the observed radar backscatter. Further analysis of the radar echo suggests that subsurface RA is ∼0.47%–0.88% if the effective size is 3 cm, which can explain the shallow sampling depth of the CE‐5 drilling device. All these results provide valuable information for understanding the geological context and properties of the returned regolith samples.

Stratigraphic Relationships in Jezero Crater, Mars: Constraints on the Timing of Fluvial‐Lacustrine Activity From Orbital Observations

AbstractOn February 18, 2021 NASA's Perseverance rover landed in Jezero crater, located at the northwestern edge of the Isidis basin on Mars. The uppermost surface of the present‐day crater floor is dominated by a distinct geologic assemblage previously referred to as the dark‐toned floor. It consists of a smooth, dark‐toned unit overlying and variably covering light‐toned, roughly eroded deposits showing evidence of discrete layers. In this study, we investigated the stratigraphic relations between materials that comprise this assemblage, the main western delta deposit, as well as isolated mesas located east of the main delta body that potentially represent delta remnants. A more detailed classification and differentiation of crater floor units in Jezero and determination of their relative ages is vital for the understanding of the geologic evolution of the crater system, and determination of the potential timeline and environments of habitability. We have investigated unit contacts using topographic profiles and DEMs as well as the distribution of small craters and fractures on the youngest portions of the crater floor. Our results indicate that at least some of the deltaic deposition in Jezero postdates emplacement of the uppermost surface of the crater floor assemblage. The inferred age of the floor assemblage can therefore help to constrain the timing of the Jezero fluviolacustrine system, wherein at least some lake activity postdates the age of the uppermost crater floor. We present hypotheses that can be tested by Perseverance and can be used to advance our knowledge of the geologic evolution of the area.

Knowledge-Driven GeoAI: Integrating Spatial Knowledge into Multi-Scale Deep Learning for Mars Crater Detection

This paper introduces a new GeoAI solution to support automated mapping of global craters on the Mars surface. Traditional crater detection algorithms suffer from the limitation of working only in a semiautomated or multi-stage manner, and most were developed to handle a specific dataset in a small subarea of Mars’ surface, hindering their transferability for global crater detection. As an alternative, we propose a GeoAI solution based on deep learning to tackle this problem effectively. Three innovative features are integrated into our object detection pipeline: (1) a feature pyramid network is leveraged to generate feature maps with rich semantics across multiple object scales; (2) prior geospatial knowledge based on the Hough transform is integrated to enable more accurate localization of potential craters; and (3) a scale-aware classifier is adopted to increase the prediction accuracy of both large and small crater instances. The results show that the proposed strategies bring a significant increase in crater detection performance than the popular Faster R-CNN model. The integration of geospatial domain knowledge into the data-driven analytics moves GeoAI research up to the next level to enable knowledge-driven GeoAI. This research can be applied to a wide variety of object detection and image analysis tasks.