多种角度比较SIFT、SURF、BRISK、ORB、FREAK算法

SIFT、SURF、BRISK、ORB、FREAK等图像特征匹配算法具有不同的鲁棒性,本文利用Mikolajczyk和Schmid所提供的数据集,对这几种算法从匹配速度,应对图像旋转变换、模糊变换、光照变换、尺度变换、视角变换时表现出的鲁棒性等角度进行比较。本文采用的评价指标是两幅图像中成功匹配的特征点对数。经比较分析后发现,SURF和FREAK算法表现出的鲁棒性较为全面。     

How to compare movement? A review of physical movement similarity measures in geographic information science and beyond

In geographic information science, a plethora of different approaches and methods is used to assess the similarity of movement. Some of these approaches term two moving objects similar if they share akin paths. Others require objects to move at similar speed and yet others consider movement similar if it occurs at the same time. We believe that a structured and comprehensive classification of movement comparison measures is missing. We argue that such a classification not only depicts the status quo of qualitative and quantitative movement analysis, but also allows for identifying those aspects of movement for which similarity measures are scarce or entirely missing.

In this review paper we, first, decompose movement into its spatial, temporal, and spatiotemporal movement parameters. A movement parameter is a physical quantity of movement, such as speed, spatial path, or temporal duration. For each of these parameters we then review qualitative and quantitative methods of how to compare movement. Thus, we provide a systematic and comprehensive classification of different movement similarity measures used in geographic information science. This classification is a valuable first step toward a GIS toolbox comprising all relevant movement comparison methods.     

IMU/DGPS辅助空三下几种布点方案的精度比较

主要介绍了在IMU/DGPS辅助空三的情况下,通过对不同的布点方案进行空三加密,对这些空三成果进行精度比较与分析,分析POS辅助空三在无地面控制点或少量的地面控制点的情况下能获得怎样的精度,能满足多大比例尺的成图要求。     

人工神经网络和决策树模型在滑坡易发性分析中的性能对比

机器学习模型广泛应用于区域性滑坡易发性分析。模型的选择关系到评价结果的可信度、准确率和稳定性。现有滑坡易发性分析模型对比研究侧重模型的预测精度。模型的稳定性和数据量敏感性对机器学习模型的性能评估同样非常重要。本文以福建省南平市蔡源流域为研究区,以四川省绵阳市北川县为验证区,从预测精度、稳定性和数据量敏感性3个方面深入对比BP（Back Propagation）人工神经网络模型和CART（Classification and Regression Tree）决策树模型在滑坡易发性分析中的效果,主要结论如下:① 在逐渐增加一定数量训练样本的过程中,BP人工神经网络模型预测精度的增长率更高。在蔡源流域内,当训练样本数量增加10 000时,BP人工神经网络模型的预测精度上升5.22%,CART决策树模型的预测精度上升2.11%。② BP人工神经网络的预测精度高于CART决策树模型,且较为稳定。在100组数据集上,BP人工神经网络模型验证集预测精度的均值和验证集滑坡样本预测精度的均值分别为81.60%和84.86%,高于CART决策树模型的72.97%和76.59%。与此同时,BP人工神经网络模型对应预测精度的标准差分别是0.32%和0.37%,小于CART决策树模型的0.35%和0.67%。③ BP人工神经网络模型分析的滑坡易发区相比CART决策树模型,更接近实际滑坡的空间分布。最后,北川县的验证实验也出现了相同的现象。     

贵州与印度的夏季风的异同分析

季风作了气象科学的一个重要课题，许多人都作了大量研究，中国和印度均处于季风区，印度季风和中国季风的研究已作了大量研究，揭露了不少事实，但中国季风的研究多指我国的东部，而贵州位于我国西南部，地处东亚季风和印度季风之间的过渡区，因此，对贵州省的论及较少，从气候的角度出发，把贵州和印度的季风的某些特征量，如降雨、风、环流等进行对比分析，在一定程度上反映了贵州和印度夏季风的主要异同。     

Effects of land‐cover changes on the hydrological response of interior Columbia River basin forested catchments

The topographically explicit distributed hydrology–soil–vegetation model (DHSVM) is used to simulate hydrological effects of changes in land cover for four catchments, ranging from 27 to 1033 km², within the Columbia River basin. Surface fluxes (stream flow and evapotranspiration) and state variables (soil moisture and snow water equivalent) corresponding to historical (1900) and current (1990) vegetation are compared. In addition a sensitivity analysis, where the catchments are covered entirely by conifers at different maturity stages, was conducted. In general, lower leaf‐area index (LAI) resulted in higher snow water equivalent, more stream flow and less evapotranspiration. Comparisons with the macroscale variable infiltration capacity (VIC) model, which parameterizes, rather than explicitly represents, topographic effects, show that runoff predicted by DHSVM is more sensitive to land‐cover changes than is runoff predicted by VIC. This is explained by model differences in soil parameters and evapotranspiration calculations, and by the more explicit representation of saturation excess in DHSVM and its higher sensitivity to LAI changes in the calculation of evapotranspiration. Copyright © 2002 John Wiley & Sons, Ltd.     

Comparison of Kuqa foreland basin with Persian Gulf Basin in salt tectonics

Compared Kuqa foreland basin with Persian Gulf Basin in development of salt layers, salt tectonics, and the relation between salt tectonics and hydrocarbon, it is concluded that the salt diapirs are relative to hydrocarbon. Searching salt diapirs and related traps in Kuqa foreland basin is important. The forming mechanism of salt tectonic in Kuqa foreland basin is different from that of Hormuz Series, but similar to that of Lower Fars Series/Gachsaran Formation. Inspired by the role of salt tectonics of Lower Fars Series/Gachsaran Formation in hydrocarbon accumulation, the authors considered that the exploration below salt layer should be enforced, and the traps below salt layer in the southern part of the Kuqa foreland basin would be found where salt layer is thicker. On the contrary, the traps should be found both above and below the salt layer in front of the northern mountain where salt layer is thin. The Triassic and Jurassic source rocks are rich in this area with great exploration prospective.     

鄂尔多斯盆地庆城地区延长组7段致密油特征及油源分析

鄂尔多斯盆地三叠系延长组7段(长7段)半深湖—深湖相优质油源岩是油气的重要来源,包括黑色页岩和暗色泥岩,作为盆地重要的供烃源岩一直是研究的热点。为了明确不同烃源岩的贡献及油气来源,通过对庆城地区长7段致密油特征进行分析,对原油进行精细划分,并对源岩有效性进行分析,进一步明确油源对应关系,研究表明,庆城地区长7段黑色页岩和暗色泥岩均对油气成藏有贡献,其中黑色页岩贡献最大,长7₃段暗色泥岩次之,长7₂和长7₁段暗色泥岩贡献较小;原油为低密度、中—低黏度和低凝固点的轻质油。根据生标物特征将原油划分为3类,A类原油具有低C₃₀重排藿烷/C₂₉藿烷、低C₂₉Ts/C₂₉藿烷、低Ts/Tm的特征,由黑色页岩贡献,主要分布于长7₁段;C类原油具有高C₃₀重排藿烷/C₂₉藿烷、高C₂₉Ts/C₂₉藿烷、高Ts/Tm的特征,主要由暗色泥岩贡献,...     

矿用钻孔超声流量自适应检测技术

为了更好地实现煤矿井下瓦斯抽采钻孔端流量监测,针对管路内介质成分复杂、纯净度差、稳定性差等因素对抽采流量参数测量准确性影响的问题,提出以时差法为测量原理,以双阈值比较法为检测方法,设计了一套矿用钻孔超声流量自适应检测系统。通过分析影响双阈值比较法测量准确性的关键问题出发,设计了检测系统的总体方案,并对其中激励信号放大电路、接收信号调理电路及由峰值检波电路与增益控制电路构成的自适应电路几项关键模块进行了详细地介绍,叙述了检测系统运行软件的工作流程。通过工况环境适应能力测试与准确度性能检验,对检测系统的功能与性能进行了检验,检验结果表明,该系统能够适应瓦斯抽采钻孔管路内的工况需求,实现高精度的气体流量测量。