基于深度学习的大型陨石坑识别方法研究

陨石坑是天体表面最为显著的地形特征,传统陨石坑识别方法主要是对小型陨石坑正负样本的二分类问题研究,且效率和精度均不高。以星体宏观视角下的大型陨石坑作为研究对象,结合图像处理和神经网络等方面的知识,创建了来自不同数据源的陨石坑样本数据库,研究了数据源对网络模型泛化能力的影响,提出了一种效率更高的陨石坑多分类识别方法。在非极大值抑制（NMS）算法基础上,提出了一种精度更高的陨石坑检测算法。经过参数优化和实验验证,构建的基于深度学习的多尺度多分类陨石坑自动识别网络框架取得了较高的准确率,在同源验证集上识别率可达0.985,在异源验证集上识别率可达0.863,并且有效改善了目标检测时检测框冗余及误检测的问题。      

Detection of sub-kilometer craters in high resolution planetary images using shape and texture features

Counting craters is a paramount tool of planetary analysis because it provides relative dating of planetary surfaces. Dating surfaces with high spatial resolution requires counting a very large number of small, sub-kilometer size craters. Exhaustive manual surveys of such craters over extensive regions are impractical, sparking interest in designing crater detection algorithms (CDAs). As a part of our effort to design a CDA, which is robust and practical for planetary research analysis, we propose a crater detection approach that utilizes both shape and texture features to identify efficiently sub-kilometer craters in high resolution panchromatic images. First, a mathematical morphology-based shape analysis is used to identify regions in an image that may contain craters; only those regions - crater candidates - are the subject of further processing. Second, image texture features in combination with the boosting ensemble supervised learning algorithm are used to accurately classify previously identified candidates into craters and non-craters. The design of the proposed CDA is described and its performance is evaluated using a high resolution image of Mars for which sub-kilometer craters have been manually identified. The overall detection rate of the proposed CDA is 81%, the branching factor is 0.14, and the overall quality factor is 72%. This performance is a significant improvement over the previous CDA based exclusively on the shape features. The combination of performance level and computational efficiency offered by this CDA makes it attractive for practical application.     

A new approach based on crater detection and matching for visual navigation in planetary landing

This paper provides an approach of crater detection and matching to visual navigation in planetary landing missions. The approach aims to detect craters on the planetary surface and match them to a landmark database during the descent phase of a planetary landing mission. Firstly an image region pairing method is proposed to detect the crater by using an image region feature detector. Then a WTA-rule is adopted to match the detected crater to the crater in database. To further reduce the false matching rate, an efficient method for reducing false matches using parameters of crater in 3-D database is proposed. Real images of planetary terrain and a semi-physical planetary landing simulation platform are utilized to test the performance of the approach, simulation results show the proposed approach is able to match the required number of craters to the database for pin-point planetary landing with a low rate of false detection and false matching, which will lead to an improved planetary landing precision.     

Single crater-aided inertial navigation for autonomous asteroid landing

In this paper, a novel crater-aided inertial navigation approach for autonomous asteroid landing mission is developed. It overcomes the major deficiencies of existing approaches in the literature, which mainly focuses on the case where craters are abundant in the camera field of view. As a result, traditional crater based methods require at least three craters to achieve crater matching, which limits their application in final landing phase where craters are scarce in the camera’s field of view. In contrast, the proposed algorithm enables single crater based crater matching based on a novel 2D-3D crater re-projection model. The re-projection model adopts inertial measurements as a reference, and re-projects the 3D crater model onto descent images to achieve the matching to its counterpart. An asteroid landing simulation toolbox is developed to validate the performance of the proposed approach. Through comparison with the state-of-the-art local image feature and crater based navigation algorithms, the proposed approach is validated to achieve a competitive performance in terms of feature matching and pose estimation accuracy with a much lighter computational cost.     

Bayesian network-based extraction of lunar impact craters from optical images and DEM data

Impact craters are among the most noticeable geomorphological features on the planetary surface and yield significant information about terrain evolution and the history of the solar system. Thus, the recognition of impact craters is an important branch of modern planetary studies. Aiming at addressing problems associated with the insufficient and inaccurate detection of lunar impact craters, a decision fusion method within the Bayesian network (BN) framework is developed in this paper to handle multi-source information from both optical images and associated digital elevation model (DEM) data. First, we implement the edge-based method for efficiently searching crater candidates which are the image patches that can potentially contain impact craters. Secondly, the multi-source representations of an impact crater derived from both optical images and DEM data are proposed and constructed to quantitatively describe the two-dimensional (2D) and three-dimensional (3D) morphology, consisting of Histogram of Oriented Gradient (HOG), Histogram of Multi-scale Slope (HMS) and Histogram of Multi-scale Aspect (HMA). Finally, a BN-based framework integrates the multi-source representations of impact craters, which can provide reductant and complementary information, for distinguishing craters from non-craters. To evaluate the effectiveness and robustness of the proposed method, experiments were conducted on three lunar scenes using both orthoimages from the Lunar Reconnaissance Orbiter (LRO) and DEM data acquired by the Lunar Orbiter Laser Altimeter (LOLA). Experimental results demonstrate that integrating optical images with DEM data significantly decreases the number of false positives compared with using optical images alone, with F₁-score of 84.8% on average. Moreover, compared with other existing fusion methods, our proposed method was quite advantageous especially for the detection of small-scale craters with diameters less than 1000 m.     

Crater detection via genetic search methods to reduce image features

Recent approaches to crater detection have been inspired by face detection’s use of gray-scale texture features. Using gray-scale texture features for supervised machine learning crater detection algorithms provides better classification of craters in planetary images than previous methods. When using Haar features it is typical to generate thousands of numerical values from each candidate crater image. This magnitude of image features to extract and consider can spell disaster when the application is an entire planetary surface. One solution is to reduce the number of features extracted and considered in order to increase accuracy as well as speed. Feature subset selection provides the operational classifiers with a concise and denoised set of features by reducing irrelevant and redundant features. Feature subset selection is known to be NP-hard. To provide an efficient suboptimal solution, four genetic algorithms are proposed to use greedy selection, weighted random selection, and simulated annealing to distinguish discriminate features from indiscriminate features. Inspired by analysis regarding the relationship between subset size and accuracy, a squeezing algorithm is presented to shrink the genetic algorithm’s chromosome cardinality during the genetic iterations. A significant increase in the classification performance of a Bayesian classifier in crater detection using image texture features is observed.     

New findings and appraisal of structural control on polygonal impact crater rim-geometry,a study from Mare Fecunditatis,Moon

Polygonal Impact Craters (PICs), having a distinct polygonal rim geometry, are common on terrestrial planets, their natural satellites such as Earth’s Moon and the asteroids. The straight segments of PIC-rims are arguably subparallel or oblique to existing fracture/fault planes in their vicinity, and such pre-existing structural weak planes are considered responsible for the shape of the PICs. The Mare Fecunditatis, a lunar maria, preserves mappable PICs as well as different geomorphic features like wrinkle ridges, grabens and pit crater chains which owe their origin to either compressional or extensional faulting. To understand the structural control, if any, on the PIC-rim geometry in Mare Fecunditatis, PICs, both simple and complex, and the deformational features are mapped, superposition relations between them are observed and trends are compared. The comparison between frequency of rim segment trends of the two types of PICs with wrinkle ridges, grabens and pit crater chains, and also statistical correlation between them, interestingly indicate that the wrinkle ridges and grabens are found to have negligible or no influence on the rim geometry of the PICs. Wrinkle ridges, known to be a group of the Mare Fecunditatis' oldest deformation features, are older than most of the preserved craters and are likely to have had control over the PIC shape. However, lack of correlation between the trends of wrinkle ridges and PIC rims indicates that most of the craters were formed after the fractures beneath the old wrinkle ridges ceased to act as mechanical discontinuity planes due to possible induration caused by fracture-filling through concomitant and later magma injection. The PICs dispersed throughout the maria could have avoided the influence of the grabens, the majority of which are located near the margin of the maria. Pit crater chains with possible deep roots and still/recently continued dike activities, were the only available weak planes which could influence the PIC rim shapes.     

Integrated method for crater detection from topography and optical images and the new PH9224GT catalogue of Phobos impact craters

In a large majority of lunar and planetary surface images, impact craters are the most abundant geological features. Therefore, it is not surprising that crater detection algorithms (CDAs) are one of the most studied subjects of image processing and analysis in lunar and planetary science. In this work we are proposing an Integrated CDA, consisting of: (1) utilization of DEM (digital elevation map)-based CDA; (2) utilization of an optical-based CDA; (3) re-projection of used datasets and crater coordinates from normal to rotated view and back; (4) correction of the brightness and contrast of a used optical image; and (5) tile generation for the optical-based CDA and an assembling of results with an elimination of multiple detections, in combination with a pyramid approach down to the resolution of the available DEM image; and (6) a final integration of the results of DEM-based and optical-based CDAs, including a removal of duplicates. The proposed CDA is applied to one specific asteroid-like body, the small Martian moon Phobos. The experimental evaluation of the proposed CDA is done by a manual verification of crater-candidates and a search for uncatalogued craters. The evaluation has shown that the proposed CDA was used successfully for cataloging Phobos craters. The major result of this paper is the PH9224GT – currently the most complete global catalogue of the 9224 Phobos craters. The possible applications of the new catalogue are: (1) age estimations for any selected location; and (2) comparison/evaluation of the different chronology and production functions for Phobos. This confirms the practical applicability of the new Integrated CDA – an additional result of this paper, which can be used in order to considerably extend the current crater catalogues.     

一种基于多尺度边缘提取的陨石坑检测算法

针对星体表面的陨石坑可用于探测器的自主导航、障碍识别等任务,提出一种基于多尺度边缘提取的陨石坑检测算法。该算法首先利用高斯金字塔得到不同尺度的陨石坑图像;其次,针对不同尺度的陨石坑图像,利用EDPF边缘提取算法对陨石坑进行边缘提取,并连接关键边缘像素点为直线段来近似表示图像边缘;然后将具有相同偏转方向的边缘直线段连接成圆弧,并将有相似半径和中心的圆弧拟合成候选圆和椭圆;最后对候选圆、椭圆进行验证。该算法的优点在于,能够准确地检测出陨石坑,有较高的检测率,且对存在较多陨石坑的图像有较好的检测结果。     

The first new application of the Mathematical Theory of Stochastic Processes to lunar and planetary science: Topography-Profile Diagrams of Mars

The Mathematical Statistics Theory (MST) and the Mathematical Theory of Stochastic Processes (MTSP) are different branches of the more general Mathematical Probability Theory (MPT) that can be used to investigate physical processes through mathematics. Each model of a stochastic process, according to MTSP, can provide one or more interpretations in the MST domain. A large body of work on impact crater statistics according to MST exists, showing cumulative crater frequency (N km⁻²) as a function of age (years) for some particular crater diameter. However, this is only one possible representation in the MST domain of the bombardment of the planetary surface modeled as a stochastic process according to MTSP. The idea that other representations are possible in the MST domain of the same stochastic process from MTSP has been recently presented. The importance of the approach is that each such mathematical-based interpretation can provide a large amount of new information. Coupled with MOLA data, Topography-Profile Diagrams (TPD) are one of the many examples that can provide a large amount of new information regarding the history of Mars. TPD consists of: (1) Topography-Profile Curve (TPC), which is a representation of the planet’s topography, (2) Density-of-Craters Curve (DCC), which represents density of craters, (3) Filtered-DCC (FDCC), which represents DCC filtered by a low-pass filter, included with the purpose of reducing the noise, and (4) Level-of-Substance-Over-Time Curve (LSOTC), which represents interpretation of the influence on the distribution of craters shown by FDCC. TPC uniquely corresponds to the computation of TPD, whereas DCC depends on algorithms for computing the elevation of each crater according to the topography, center coordinates, and radius of impact crater, and FDCC relies on the architecture of the custom designed low-pass filter for filtering DCC. However, all variations of DCC and FDCC, which includes the various impact crater data sets, showed a correlation among the density of craters and elevation over 70–80% of the planet surface. Additionally, if we assume that the ocean primarily caused the noted correlation, LSOTC offers a mathematical approach for estimating topographic change of the ocean’s extent over time. Accordingly, TPD is the first new practical application of MTSP to lunar and planetary sciences, showing correlation of topography to a physical process.     

Hybrid method for crater detection based on topography reconstruction from optical images and the new LU78287GT catalogue of Lunar impact craters

Impact craters are ubiquitous and well-studied structures of high geological relevance on the surfaces of the Earth’s Moon, the terrestrial planets, the asteroids and the satellites of the outer planets. Therefore, it is not surprising that crater detection algorithms (CDAs) are one of the most studied subjects of image processing and analysis in lunar and planetary science. In this paper we are proposing a Hybrid CDA: a modified DEM (digital elevation map) reconstruction method used as a step in an existing CDA based on Hough transform. The new Hybrid CDA consists of: (1) reconstruction of topography from optical images using a shape from shading approach; (2) utilization of the DEM-based CDA; (3) correction of brightness and contrast of optical images used in order to be more suitable for evaluation of detections. An additional result of this work is a new method for evaluation of topography reconstruction algorithms, using a DEM-based CDA and an earlier approach for evaluation of CDAs. The new Hybrid CDA was tested using two Chandrayaan-1 Moon Mineralogy Mapper (M³) images and two excerpts of the Lunar Reconnaissance Orbiter (LRO) Wide Angle Camera (WAC) global optical image mosaic. As a result, the number of craters inside these four regions increased considerably from 1754 (as available in the previous LU60645GT catalogue) to 19 396 craters (as available in the resulting new LU78287GT catalogue). This confirmed the practical applicability of the new Hybrid CDA, which can be used in order to considerably extend current crater catalogues.     

一种基于自动特征学习的陨石坑区域检测算法

基于陨石坑的视觉导航技术成为一种新颖的高精度空间探测自主导航方式,如何从导航图像中精确地提取陨石坑区域是实现基于陨石坑视觉导航的首要条件。针对这一问题,根据陨石坑导航图像特点,提出了一种基于自动特征学习的陨石坑区域检测算法。首先,基于最大稳定极值区域检测算法提取陨石坑候选区域;其次,利用卷积神经网络（CNN）自动学习提取候选区域的特征;最后,通过支持向量机（SVM）实现候选区域的精确分类,得到真实的陨石坑区域。大量的仿真实验表明:与传统的基于人工特征的陨石坑区域检测算法相比,提出的基于自动特征学习的陨石坑区域检测算法具有更高的检测精度和更好的鲁棒性,在通用火星表面陨石坑数据集上,所提算法的F₁度量指标较于传统算法高出8%,可以广泛地应用于基于陨石坑的视觉导航算法中的陨石坑区域提取,为基于陨石坑视觉导航算法提供精确的导航路标输入。      

月球撞击坑辐射纹的形成和消失机制及可能的年代学应用

辐射纹是年轻撞击坑周缘呈辐射状分布的明亮细条纹,是月表最显著的地貌特征之一,也是月球科学领域热点研究之一.根据目前对于月球撞击坑辐射纹的形成和类型的认识,分析辐射纹消失相关的空间风化、撞击导致的物质混合等地质过程;比较不同形成年龄撞击坑辐射纹的光学成熟度(OMAT)剖面,发现溅射物逐渐成熟过程中OMAT剖面的演变过程,...     

Application of hydrocode modelling to the study of hypervelocity impact crater morphology

In order to obtain a better understanding and model of the natural and artificial particulate environment from measurements of impact damage features on returned spacecraft materials, it is necessary to be able to determine how the size and shape of an impact feature are related to the parameters of the impacting particle. The AUTODYN-3D hydrocode has been used to study the effects of projectile density, velocity and impact angle on the depth, diameter and ellipticity of the impact craters. The results are used to determine the distributions of crater depth to crater diameter ratios and of crater ellipticities to be expected on an aluminium surface exposed to an isotropic distribution of incident particles of given densities and velocities. Comparison of these calculated distributions with those observed for craters on aluminium clamps on various faces of the Long Duration Exposure Facility shows that particles with a wide range of densities, including significant proportions both greater and smaller than that of aluminium, were responsible for these craters.     

小行星形貌测绘与表征技术

小行星表面形貌测绘是深空小行星探测的首要任务。提出了一种适应探测器抵近观测的立体视觉在线测绘小行星形貌的方法,即先由立体相机获得所摄重叠区的三维地形,再用前后立体模型的连接点将各个独立模型连成一个完整地形信息,经整体最小二乘平差,确定小行星的形貌模型及特征;同时提出了一种基于等值线分析的撞击坑特征提取方法,即通过提取、分析地形等值线识别出地形中撞击坑特征。实验结果显示,本文所构建的原型系统能够快速重建出探测区域的三维地形,并识别出地形中撞击坑特征,证明了所提方法具有实用性。     

Synergetic use of SAR and Thermal Infrared data to study the physical properties of the lunar surface

The surface layer of the Moon preserves vital evidences of lunar impact and cratering processes due to the absence of any Aeolian and fluvial erosion processes acting on it. By examining these evidences, which are recorded throughout the evolutionary history of the Moon, several basic aspects of lunar science can be understood, and this has direct relevance to the surfaces of other airless bodies within the solar system. In this study, rock abundance data obtained from Thermal Infrared (TIR) observations and radar Circular Polarization Ratio (CPR) data sets obtained from polarimetric SAR observations were correlated at some sample sites on the lunar surface. Preliminary results yielded qualitative and quantitative estimates for surface rock abundances. Except at distal ejecta deposits of young, bright craters a general correlation was observed between the two datasets. Mixed results were observed from the impact melt flows where the situation is complex due to the possible subsurface-volume and volume-subsurface interactions of the radar waves. But the flow features were clearly separated from the interior and ejecta regions of their parent craters in terms of CPR and rock abundances. The extent and distributions of pyroclastic deposits and dark haloed regions could not be distinctly identified at the resolution of datasets utilized. Near Gerasimovich D crater, the Diviner Radiometer has provided the first TIR observations of a newly discovered impact melt flow which was not visible in the optical imagery. This facilitated the first ever quantitative comparisons of the radar CPR and rock abundance values near such a region. Also, significant differences in spatial patterns between the radar and rock concentration data sets were observed, owing to the differences in the sensitivity of the two observations.     

Hydrocode modelling of hypervelocity impacts on ice

Experimental data are now widely available for the size of craters resulting from hypervelocity impacts of millimetre scale projectiles onto water ice targets. At such size scales the bowl shaped crater formed in ductile materials, or in larger scale impacts, is here surrounded by a large spallation zone due to the brittle nature of the ice. Modelling of these impacts therefore has to take account of this spallation. Here we used the iSALE2 hydrocode to simulate such impacts and compared the results to experimental data. We found that it was possible to reproduce the experimental data over a range of speeds (1–7 km s⁻¹) for aluminium and copper projectiles. Initially, to reproduce the large spallation regions around the craters it was assumed that above a certain degree of damage, material was removed by spallation. However this simple one-parameter model failed to model the crater depth adequately. Accordingly, to obtain the best agreement of the simulations with the experimental data, a two-step ice strength was introduced, whereby above a critical amount of damage (0.95), the yield strength reduced from 1 MPa (intact) to 70 kPa (damaged). As a result, experimental data for crater depth and diameter and the results of the simulations agree to within 6% for diameter and 5% for depth over the impact energy range used in the experiments (1–240 J).     

Exploring links between crater floor mineralogy and layered lunar crust

Floors of similar sized craters, representing material from similar depth horizons, have been studied to explore their suitability as mineralogy indicators at various depths within the lunar crust. Clementine UV–vis multispectral data was used to generate mineral abundance maps of crater floors and surroundings using a modified version of algorithm given by Pieters et al. (2001) [Pieters, C.M., Head, J.W., Gaddis, L., Jolliff, B. and Duke, M. Rock types of the south pole aitken basin and extent of basaltic volcanism, JGR (106) E11, 28001–28022, 2001.]. Substantial processing of the crater floor material due to variety of geological processes is evident in the generated maps, making straight forward interpretations difficult. However, systematic compositional trends in fresh craters on the floors of target craters seem to indicate the feasibility of such an effort.     

Analysis of the brightness temperature features of the lunar surface using 37 GHz channel data from the Chang'E-2 microwave radiometer

Compared with other remote observations, brightness temperatures (T_B) derived from microwave emission measurements provide a unique means to characterize the physical properties of the lunar surface. Using Chang’E-2 microwave radiometer data, we produced 12 global T_B images of the lunar surface during a diurnal cycle with different local times separated by approximately 2?h. There are two types of remarkable T_B units on the lunar surface, the “hot regions” occurring during the lunar day in the lunar Maria regions and the “microwave cold spots” occurring during the nighttime typically related to young craters. Compared with their surroundings, the hot regions are much warmer during the lunar day and slightly colder at night, while the microwave cold spots are much colder during the lunar night and slightly warmer in the daytime. Moreover, the T_B heating and cooling rates of these two units are larger than others at the same average latitude where they are located during the lunar day, especially after sunrise and before sunset. The hot regions have a good agreement with the mare regions with high TiO₂ abundance. Besides, brightness temperatures in the lunar Maria correlate closely with their TiO₂ abundance. For most microwave cold spots, they agree with the young craters, and their brightness temperature distributions have a significant negative correlation with the lunar surface nighttime temperature and rock abundance.     

基于LOLA数据的冯·卡门撞击坑太阳辐射研究

冯·卡门（Von Kármán）撞击坑是“嫦娥4号”的候选着陆区之一。基于LOLA高程数据,对当前的月球光照模型做出改进,建立了月表太阳辐射模型,对冯·卡门地区2018年太阳辐射进行了数值模拟分析。结果表明：地形对太阳辐射的影响很大,撞击坑的南部坑壁、中央峰北部以及内部小撞击坑南部坑壁接收的太阳辐射能较多,坑底平原大部分地区接收的太阳辐射能在（0.9~1）×10¹⁰ J/m²之间;不考虑月面坡度时,太阳辐射能量主要受纬度的影响,计算区域的变化范围为（0.87~1.01）×10¹⁰ J/m²。结合月表坡度和光照条件提出了两个候选着陆区（S1区和S2区）：S1区位于坑底南部平原,地势更平缓,日出更早,光照时间更长;S2区位于中央峰西北侧,接收的太阳辐射能量更多。两区全年平均接收的太阳辐射能分别为9.31×10⁹ J/m²和9.65×10⁹ J/m²,7月份光照时间最长,更适宜着陆。