遥感影像智能解译样本库现状与研究

我国遥感对地观测等项目顺利实施,获取了大量时效性强、覆盖范围广、信息量丰富的遥感数据.但遥感影像智能化自动处理技术发展仍相对滞后,无法满足区域/全球大范围地物信息快速提取的需求.近年来,人们利用深度学习技术显著提高了影像特征提取成效,但由于所使用的深度学习样本数量和类型有限,对于多源遥感影像的自动解译能力仍然不足.本文...     

Mortality by cause of death in Colombia: a local analysis using spatial econometrics

Journal of Geographical Systems - Colombia is undergoing major changes in mortality patterns. National- and department-level cause-specific analyses have previously been carried out, but very...     

Multi‐Scale Qualitative Location: A Topology‐Based Model

Qualitative locations describe spatial objects by relating the spatial objects to a frame of reference (e.g. a regional partition in this study) with qualitative relations. Existing models only formalize spatial objects, frames of reference, and their relations at one scale, thus limiting their applicability in representing location changes of spatial objects across scales. A topology‐based, multi‐scale qualitative location model is proposed to represent the associations of multiple representations of the same objects with respect to the frames of reference at different levels. Multi‐scale regional partitions are first presented to be the frames of reference at multiple levels of scale. Multi‐scale locations are then formalized to relate multiple representations of the same objects to the multiple frames of reference by topological relations. Since spatial objects, frames of reference, and topological relations in qualitative locations are scale dependent, scale transformation approaches are presented to derive possible coarse locations from detailed locations by incorporating polygon merging, polygon‐to‐line and polygon‐to‐point operators.     

自适应高斯滤波与SFIM模型相结合的全色多光谱影像融合方法

全色-多光谱影像融合技术可以显著提高遥感影像的地物判别能力,但是空间信息融入度与光谱信息保真度是相互矛盾的一组性质,一般方法往往无法平衡这两方面。SFIM算法具有良好的光谱信息保持能力,但是其空间信息融入度较差,影响了整体的融合效果。为此,本文分析了SFIM模型的原理与特点,提出一种自适应高斯滤波与SFIM模型相结合的全色多光谱影像融合方法（AGSFIM）。以均值调整后的多光谱整体平均梯度为标准来计算高斯滤波的最优参数,将下采样全色影像的清晰度调整至同样水平,以保证融合结果的空间信息融入度与光谱信息保真度之间的平衡。利用6种融合算法对“北京二号”（Beijing-2）、“资源三号”02星（ZY-3 02）数据进行对比试验,表明在良好的光谱保持能力的前提下,改进方法可以有效克服SFIM算法空间信息融入不足的缺点。     

临沂市地质灾害预警预报与指挥系统的设计与实现

本文结合国内外有关地质灾害预警预报的相关理论研究,论述了临沂市地质灾害预警预报与指挥系统的总体框架和功能设计,介绍了本系统的预警预报、应急指挥、态势标绘等功能。系统旨在逐步提高临沂市的地质灾害预警预报技术水平,为各级政府制定国民发展计划和城镇规划提供重要的科学依据,为减灾防灾提供技术支撑。     

Modeling Urban Land Use Changes Using Support Vector Machines

Support Vector Machines (SVM) is a machine learning (ML) algorithm commonly applied to the classification of remotely sensing data and more recently for modeling land use changes. However, in most geospatial applications the current literature does not elaborate on specifications of the SVM method with respect to data sampling, attribute selection and optimal parameters choices. Therefore the main objective of this study is to present and investigate the SVM technique for modeling urban land use change. The SVM model building procedure is presented together with the detailed evaluation of the output results with respect to the choice of datasets, attributes and the change of SVM parameters. Geospatial datasets containing nine land use classes and spatial attributes for the Municipality of Zemun, Republic of Serbia were used for years 2001, 2003, 2007 and 2011. The Correlation‐based Feature Subset method, kappa coefficient, Area Under Receiver Operating Characteristic Curve (AUC) and kappa simulation were used to perform the model evaluation and compare the model outputs with the real land use datasets. The obtained results indicate that the SVM‐based models perform better when implementing balanced data sampling, reduced data sets to informative subsets of attributes and properly identify the optimal learning parameters.     

A hybrid approach to modeling territorial control in violent armed conflicts

Territorial control is central to the understanding of violent armed conflicts, yet reliable and valid measures of this concept do not exist. We argue that geospatial analysis provides an important perspective to measure the concept. In particular, measuring territorial control can be seen as an application of calculating service areas around points of control. The modeling challenge is acute for areas with limited road infrastructure, where no complete network is available to perform the analysis, and movements largely occur off road. We present a new geospatial approach that applies network analysis on a hybrid transportation network with both actual road data and hexagon‐fishnet‐based artificial road data representing on‐road and off‐road movements, respectively. Movement speed or restriction can be readily adjusted using various input data. Simulating off‐road movement with hexagon‐fishnet‐based artificial road data has a number of advantages including scalability to small or large study areas and flexibility to allow all‐directional travel. We apply this method to measuring territorial control of armed groups in Sub‐Saharan Africa where inferior transport infrastructure is the norm. Based on the Uppsala Conflict Data Program's (UCDP) Georeferenced Event Data (GED) as well as spatial data on terrain, population locations, and limited transportation networks, we enhance the delineation of the specific areas directly controlled by each warring party during civil wars within a given travel time.     

基于CPU和GPU协同处理的光学卫星遥感影像正射校正方法

本文系统地探讨了基于CPU和GPU协同处理的光学卫星遥感影像正射校正方法。首先使用“层次性分块”策略设计了基于CPU和GPU协同处理的正射校正方法,然后通过配置选择优化和存储层次性访问等手段进一步提高了方法执行效率。在Tesla M2050 GPU上对资源三号卫星下视全色影像进行正射校正的实验结果表明,本文方法大幅提高了光学卫星遥感影像正射校正效率,与传统串行正射校正算法相比,加速比最高达到110倍以上,相应的处理时间压缩至5s以内,可满足对大数据量光学卫星遥感影像进行快速正射校正的要求。