New GOES imager algorithms for cloud and active fire detection and fire radiative power assessment across North, South and Central America

Vegetation fires are a key global terrestrial disturbance factor and a major source of atmospheric trace gases and aerosols. Therefore, many earth-system science and operational monitoring applications require access to repetitive, frequent and well-characterized information on fire emissions source strengths. Geostationary imagers offer important temporal advantages when studying rapidly changing phenomena such as vegetation fires. Here we present a new algorithm for detecting and characterising active fires burning within the imager footprints of the Geostationary Operational Environmental Satellites (GOES), including consideration of cloud-cover and calculation of fire radiative power (FRP), a metric shown to be strongly related to fuel consumption and smoke emission rates. The approach is based on a set of algorithms now delivering near real time (NRT) operational FRP products from the Meteosat Spinning Enhanced Visible and Infra-Red Imager (SEVIRI) imager (available from http://landsaf.meteo.pt/), and the GOES processing chain presented here is designed to deliver a compatible fire product to complete geostationary coverage of the Western hemisphere. Results from the two GOES imagers are intercompared, and are independently verified against the well regarded MODIS cloud mask and active fire products. We find that the detection of cloud and active fires from GOES matches that of MODIS very well for fire pixels having FRP > 30 MW, when the GOES omission error falls to less than 10%. The FRP of fire clusters detected near simultaneously by both GOES and MODIS have a bias of only 22 MW, and a similar bias is found when comparing near-simultaneous GOES East and GOES West FRP observations. However, many fire pixels having FRP < 30 MW remain undetected by GOES, probably unavoidably since it has a much coarser spatial resolution than MODIS. Adjustment using data from the less frequent but more accurate views obtained from high spatial resolution polar orbiting imagers could be used to bias correct regional FRP totals. Temporal integration of the GOES FRP record indicates that during the summer months, biomass burning combusts thousands of millions of tonnes of fuel daily across the Americas. Comparison of these results to those of the Global Fire Emissions Database (GFEDv2) indicate strong linear relationships (r² > 0.9), suggesting that the timely FRP data available from a GOES real-time data feed is likely to be a suitable fire emissions source strength term for inclusion in schemes aiming to forecast the concentrations of atmospheric constituents affected by biomass burning.     

基于Hadoop的云存储的研究及实现

首先描述了云计算、云存储的概念、云计算的体系结构及云存储的架构模式;然后介绍了Hadoop工作原理及其文件存储的方法;最后基于eyeOS的Web操作系统,对传统的文件存储方法进行改进,采用Hadoop的HDFS技术实现文件的分布式存储及容错控制。     

江西电网备用调度交换网组网研究

文章详细介绍了江西电网调度交换网的现状,论述了省级备调系统的建设方案及原则,提出了备用调度交换系统单独组网和合并组网两种方案的特点,并对建设中的注意事项进行阐述。同时对上述两种组网方案进行了可靠性计算。阐明了两种组网方式均大幅提高了整个调度通信网的可靠性,为目前省级备用调度交换系统建设提出了更可靠的冗余解决方案。     

基于SOA技术的高校URP建设的研究

通过基于SOA技术的高校管理工作一体化信息系统软件的设计,综合性地量化学校各项重要信息,使信息一体化程度更为集中,帮助管理工作者、校务管理人员做出动态的管理决策调整,对于URP流程改造提供具体的实践模型参考。同时研究出一种能够应用于各类普通高校管理工作的无线与网络相结合的软件系统设计模型,提供一种可量化的工作数据系统样本,以比较低廉的代价解决校园信息化过程中的信息孤岛问题,进一步提升校园信息一体化程度。     

最大似然估计ROC凸包曲线算法

采用列联表表示特定阈值下入侵检测系统(IDS)的性能,使用ROC曲线对不同阈值下IDS的总体性能进行评估。在实际应用中,基于样本数据集,通过计算几何的方式得到ROC凸包曲线(ROCCH)。为降低ROCCH的计算复杂度,在曲线下面积(AUC)最大似然估计的条件下,通过保序回归得到最大似然估计ROC曲线(MLE-ROC)算法。实验表明,MLE-ROC算法在降低计算复杂度的同时提高了AUC的近似程度。     

基于OpenCL的MD5破解算法

在基于GPU的异构平台上,采用开放计算语言(OpenCL)实现破解算法,利用分轮生成攻击密码、图形渲染管线加速存取以及多密码并行等方法对算法进行优化,在Intel四核CPU Q8230(2.3 GHz)和一片NVIDIA GT200组成的平台上进行实验。实验结果表明,在相同CPU平台上该算法能够获得高于破解软件John the ripper 17倍的破解速度。     

基于WinCC的整备作业安全监控系统

为实现电力机车整备作业安全监控自动化,结合整备作业现场控制要求,提出基于视窗控制中心(WinCC)组态软件的电力机车整备作业安全监控系统。采用组态软件的模块化设计方法,利用网络通信、监控和现场总线技术,实现以WinCC为平台的人机界面和以S7-300PLC为主站、S7-200PLC为从站的PROFIBUS-DP总线网络控制系统。测试结果表明,该系统可实现多股道同时作业,并实时实现三级分层间的数据交换。     

基于ESCA系统的层次化显式访存机制研究

针对高性能混合计算系统中的存储墙问题,在分析其计算模式特点及传统访存机制局限性的基础上,提出适用于混合计算系统的层次化显式存储访问机制,并基于ESCA多核处理器系统进行实现和评测。实验结果显示,针对核心应用程序DGEMM,延迟隐藏能够占据整体运行时间的56%,并获得1.5倍的加速比,能弥补计算与存储访问间的速度差异,提高系统计算效率。     

东营云计算系统架构关键技术的研究

云计算是一种新的用户体验和业务模式,它是在分布式计算的基础上发展而来的,黄河三角洲云计算中心是东营创建智慧城市的积极尝试。介绍云计算的概念、特点和云计算的系统架构,以IBM的云计算技术为例,结合黄河三角洲云计算中心,研究云平台系统架构的关键技术,指出了云计算的发展趋势。     

Annotation-based access control for collaborative information spaces

Web 2.0 social platforms (e.g., Flickr, YouTube) and Collaborative Working Environments (e.g., Microsoft SharePoint, BSCW) provide Web-based collaborative information spaces which enable common users and/or professionals to work together and share their online resources. Most of these collaborative information spaces provide role-based or group-based, coarse-grained access control policies which cannot successfully cope with the requirements posed by massive and open collaboration. In this paper, we present an annotation-based access control (AnBAC) model supported by a Collaboration Vocabulary (CoVoc) as a more flexible and user-centric access control approach. Based on this, we developed two tools: Uncle-Share is a gadget that provides annotation-based access control and can be equipped with CoVoc for annotating collaborative relationships. Who-With-Whom uses CoVoc to visualize extended social networks in order to help users to select appropriate contacts to grant access to resources.