An evaluation of Guided Regularized Random Forest for classification and regression tasks in remote sensing

New Earth observation missions and technologies are delivering large amounts of data. Processing this data requires developing and evaluating novel dimensionality reduction approaches to identify the most informative features for classification and regression tasks. Here we present an exhaustive evaluation of Guided Regularized Random Forest (GRRF), a feature selection method based on Random Forest. GRRF does not require fixing a priori the number of features to be selected or setting a threshold of the feature importance. Moreover, the use of regularization ensures that features selected by GRRF are non-redundant and representative. Our experiments based on various kinds of remote sensing images, show that GRRF selected features provides similar results to those obtained when using all the available features. However, the comparison between GRRF and standard random forest features shows substantial differences: in classification, the mean overall accuracy increases by almost 6% and, in regression, the decrease in RMSE almost reaches 2%. These results demonstrate the potential of GRRF for remote sensing image classification and regression. Especially in the context of increasingly large geodatabases that challenge the application of traditional methods.     

烟台市地质灾害发育现状及影响因素分析

2016年山东省开展各市地质灾害排查工作,烟台市作为地质灾害较发育区,是该次排查重点工作区之一。该文以现状调查为基础,查明了烟台市共发育471处地质灾害隐患点,地质灾害规模以小型为主,主要类型为崩塌、滑坡、泥石流、采空塌陷和岩溶塌陷,其中崩塌隐患点占地质灾害总数的68%,数量最多,其次为泥石流和采空塌陷,各占总数的15%和14%,滑坡数量最少,仅占总数的1%,并介绍了各类地质灾害分布特征。在此基础上,对地质灾害的主要影响因素——地形地貌、岩土体类型、地质构造、降水和人类工程活动等进行了分析。最后提出了在进行工程建设活动的同时重视其引发的地质灾害问题,以减少和预防其带来的次生地质灾害。     

波段选择协同表达高光谱异常探测算法

高光谱遥感影像具有光谱分辨率极高的特点,承载了大量可区分不同类型地物的诊断性光谱信息以及区分亚类相似地物之间细微差别的光谱信息,在目标探测领域具有独特的优势。与此同时,高光谱遥感影像也带来了数据维数高、邻近波段之间存在大量冗余信息的问题,高维度的数据结构往往使得高光谱影像异常目标类和背景类之间的可分性降低。为了缓解上述问题,本文提出了一种基于波段选择的协同表达高光谱异常探测算法。首先,使用最优聚类框架对高光谱波段进行选择,获得一组波段子集来表示原有的全部波段,使得高光谱影像异常目标类与背景类之间的可分性增强。然后使用协同表达对影像上的像元进行重建,由于异常目标类和背景类之间的可分性增强,对异常目标像元进行协同表达时将会得到更大的残差,异常目标像元的输出值增大,可以更好地实现异常目标和背景类的分离。本文使用了3组高光谱影像数据进行异常目标探测实验,实验结果表明,该方法与其他现有高光谱异常目标探测算法对比,曲线下面积AUC(Area Under Curve)值更高,可以更好地实现异常目标与背景分离,能够更有效地对高光谱影像进行异常目标探测。     

Outflow channel sources, reactivation, and chaos formation, Xanthe Terra, Mars

The undulating, warped, and densely fractured surfaces of highland regions east of Valles Marineris (located north of the eastern Aureum Chaos, east of the Hydraotes Chaos, and south of the Hydaspis Chaos) resulted from extensional surface warping related to ground subsidence, caused when pressurized water confined in subterranean caverns was released to the surface. Water emanations formed crater lakes and resulted in channeling episodes involved in the excavation of Ares, Tiu, and Simud Valles of the eastern part of the circum-Chryse outflow channel system. Progressive surface subsidence and associated reduction of the subsurface cavernous volume, and/or episodes of magmatic-driven activity, led to increases of the hydrostatic pressure, resulting in reactivation of both catastrophic and non-catastrophic outflow activity. Ancient cratered highland and basin materials that underwent large-scale subsidence grade into densely fractured terrains. Collapse of rock materials in these regions resulted in the formation of chaotic terrains, which occur in and near the headwaters of the eastern circum-Chryse outflow channels. The deepest chaotic terrain in the Hydaspis Chaos region resulted from the collapse of pre-existing outflow channel floors. The release of volatiles and related collapse may have included water emanations not necessarily linked to catastrophic outflow. Basal warming related to dike intrusions, thermokarst activity involving wet sediments and/or dissected ice-enriched country rock, permafrost exposed to the atmosphere by extensional tectonism and channel incision, and/or the injection of water into porous floor material, may have enhanced outflow channel floor instability and subsequent collapse. In addition to the possible genetic linkage to outflow channel development dating back to at least the Late Noachian, clear disruption of impact craters with pristine ejecta blankets and rims, as well as preservation of fine tectonic fabrics, suggest that plateau subsidence and chaos formation may have continued well into the Amazonian Period. The geologic and paleohydrologic histories presented here have important implications, as new mechanisms for outflow channel formation and other fluvial activity are described, and new reactivation mechanisms are proposed for the origin of chaotic terrain as contributors to flooding. Detailed geomorphic analysis indicates that subterranean caverns may have been exposed during chaos formation, and thus chaotic terrains mark prime locations for future geologic, hydrologic, and possible astrobiologic exploration.     

The FAST/SKA site selection in Guizhou province

Many karst depressions with diameters of 300 m to 500 m, suitable for constructing Arecibo-style radio telescopes, were identified in the south of Guizhou Province by Remote Sensing (RS) and Geographic Information System (GIS) technologies together with field investigations. Fundamental topography and landform databases were established for 391candidate depressions, and using GIS the3-dimensional images of depressions, at a scale of 1:10000, were then simulated to fit a spherical antenna. This revised version was published online in July 2006 with corrections to the Cover Date.     

辽宁省深部构造特征及其地质意义

本文应用了最新取得的“五统一”区域重力成果，并经过对浅表松散槽（如下辽河断陷，大岩体等）的密度亏损进行补偿改正后，求取了全省的区域重力场及其垂向二次导数，计算了莫氏面等深度图。在分析区域重力场及其垂向二次导数异常特征和莫氏面起伏特征的基础上，划分出辽宁深部构造的基本格局。从平面上探讨了深部构造与地形地势、地质构造及矿产分布的关系，从剖面上剖析了辽宁省地壳分层的宏观特征     

A stepwise method for determining the number of component distributions in a mixture

This paper considers the problem of estimatingm, the number of components in a finite mixture of distributions from a parametric family. A step-up procedure using the bootstrap method is proposed. Some properties of the procedure are illustrated with simulation studies. An example of the method, applied to orientation of beach clasts, is given.     

菏泽市地质灾害现状与防治

菏泽市地质灾害类型主要为地震、地面塌陷、地面沉降、地裂缝、砂土液化、土地盐渍化及地氟病等，其发育与分布受地质构造、水文地质条件的控制以及气象、人类活动等因素的影响。就其分布、危害、致灾原因等做了分析，并提出了防治建议和对策。     

Screening and ranking framework for geologic CO<Subscript>2</Subscript> storage site selection on the basis of health,safety, and environmental risk

A screening and ranking framework (SRF) has been developed to evaluate potential geologic carbon dioxide (CO₂) storage sites on the basis of health, safety, and environmental (HSE) risk arising from CO₂ leakage. The approach is based on the assumption that CO₂ leakage risk is dependent on three basic characteristics of a geologic CO₂ storage site: (1) the potential for primary containment by the target formation; (2) the potential for secondary containment if the primary formation leaks; and (3) the potential for attenuation and dispersion of leaking CO₂ if the primary formation leaks and secondary containment fails. The framework is implemented in a spreadsheet in which users enter numerical scores representing expert opinions or published information along with estimates of uncertainty. Applications to three sites in California demonstrate the approach. Refinements and extensions are possible through the use of more detailed data or model results in place of property proxies.     

Rain-induced debris and mudflow triggering factors assessment in the Santiago cordilleran foothills,Central Chile

Debris and mudflows are some of the main geological hazards in the mountain foothills of the Chilean capital city of Santiago. There, the risk of flows triggered in the basins of ravines that drain the range into the city increases with time due to the city growth. A multivariate statistical study based on the logistic regression method is presented. The model provides equations that allow the computation of combined meteorological triggering factors associated with a probability of rain-induced flow occurrence. Daily rainfall, accumulated rainfall and the snowfall level, traditionally considered as the relevant factors, are analysed for a 25-year period. The results show a strong relevance of the rainfall on the day of the flow event over the other factors. However, the relatively low probabilities returned for some real flow events suggest that the model does not capture all the significant variables and the problem is more complex than as it has been traditionally assumed, and further investigations are needed to develop predictive models of flow triggering.