基于改进U-Net网络与联合损失函数的海南自然保护区高分辨率遥感变化检测模型

海南自然保护区遥感监测对森林资源监测、生态保护及热带亚热带地表研究具有重要意义。基于深度学习方法,针对海南省自然保护区大范围变化检测问题,对U-Net网络结构进行了改进,在每一个卷积层后加入标准化层,以跳线连接的形式将原有卷积模块改进为优化模块,同时在编码器底端添加金字塔池化模块以更好提取全局信息,形成了改进U-Net网络模型。模型训练采用基于交叉熵损失函数和广义骰子损失函数构建的联合损失函数,配合多种优化策略实现端到端的地物变化信息提取。该模型应用于公开数据集和研究构建的海南自然保护区数据集的变化检测任务,总体精度分别为97.21%(Kappa系数0.88)和95.12%(Kappa系数0.90),相比原始U-Net效果提升显著。     

一种改进的CSCA+PCA化工过程异常检测技术

为了解决化工过程异常检测时因参数众多且数据庞杂而导致一些异常无法被有效检出的问题,在Brownlee的克隆选择分类算法（CSCA）基础上,通过引入主成分分析（PCA）技术,进行数据降维和数据重整,探讨了人工免疫算法在化工过程异常检测中的适用效果和技术方案,以TE过程数据作为样本进行异常检测和分类实验。结果表明,过程异常数据的规模、属性的数目对CSCA异常检测效果具有明显影响,而通过主成分分析进行数据降维之后,CSCA检测效果有所提高;进一步的数据重整之后,CSCA对过程异常分类辨识的准确率可提升到85%以上;基于CSCA+PCA的数据降维及重构之后的过程异常检测技术方案,可以获得较高的异常检测准确率,从而一定程度上为化工过程安全运行提供技术保障。     

A Method for Designing Detection Monitoring Networks in Layered Aquifers with Non-uniform Flow

An integer programming method was devised to locate detection monitoring wells in layered aquifers. The method is applicable to aquifers with non-uniform groundwater flow, and it does not require that a compliance boundary be linear or perpendicular to the direction(s) of groundwater flow. In each layer, monitoring sites are defined along curvilinear transects that parallel equipotential lines. The model can be formulated to allocate wells to the transect with the highest detection efficiency, or to establish multiple lines of defense against contaminant migrating to a compliance boundary. Detection efficiencies of alternative monitoring transects are calculated from parameters obtained via numerical modeling of contaminant transport. These parameters include the narrowest plume that could traverse a monitoring transect, and the zone width of potential contaminant migration at the transect. Problem formulations are compact, and computational requirements are low relative to alternative approaches for designing detection monitoring networks in aquifers. An application to a glacial outwash aquifer demonstrates the utility of the method.     

HPF(高通数字滤波器)的电网谐波检测技术的研究与应用

为有效地消除电力谐波污染、确保电网运行安全,必须采用有源电力滤波技术;为改善有源电力滤波器的性能,必须采取先进的谐波检测方法、高效的逆变技术。笔者在研究有源电力滤波技术的基础上,对有源电力滤波的关键技术即谐波检测与分离技术,进行了探讨和深入研究,给出了一种HPF(高通数字滤波器)的谐波检测与分离方法,进行了理论验算,利用MATLAB仿真结果对该检测技术进行实验研究。仿真和实验结果表明:该法比常用的LPF(低通数字滤波器LowPassDigitalFilter)的谐波检测分离方法具有更好的检测实时性;在电网电压有畸变时也能达到很高的检测精度;该检测技术应用于混合有源电力滤波器后能大大改善滤波性能。     

USING UNCENSORED TRACE-LEVEL MEASUREMENTS TO DETECT TRENDS IN GROUND WATER CONTAMINATION1

ABSTRACT: Environmental decision making involving trace-levels of contaminants can be complicated by censoring, the practice of reporting concentrations either as less than the limit of detection (LOD) or as not detected (ND) when a test result is less than the LOD. Censoring can result in data series that are difficult to meaningfully summarize, graph, and analyze through traditional statistical methods. In spite of the relatively large measurement errors associated with test results below the LOD, simple and meaningful analyses can be carried out that provide valuable information not available if data are censored. For example, an indication of increasing levels of contamination at the fringe of a plume can act as an early warning signal to trigger further study, an increased sampling frequency, or a higher level of remediation at the source. This paper involves the application of nonparametric trend analyses to uncensored trace-level groundwater monitoring data collected between March 1991 and August 1994 on dissolved arsenic and chromium for seven wells at an industrial site in New York.     

NONPARAMETRIC STATISTICAL METHODS IN URBAN HYDROLOGIC RESEARCH1

ABSTRACT. In urban hydrologic studies, it is often necessary to determine the effect of changes in urban land use patterns on such runoff characteristics as flood peaks and flow volumes. Nonparametric statistical methods have certain properties that make them a valuable tool for detecting hydrologic change caused by a treatment, such as urbanization, that changes watershed over a period of time. As many hydrologists do not have a working familiarity with nonparametric methods, a number of them are used for illustrative purposes to analyze the effect of urbanization on 24 years of annual flood peaks for a Louisville, Kentucky, watershed. In the example, urbanization was found to increase the central tendency, but not the dispersion of the peaks. Peak flows modeled by holding watershed parameters constant were also found to be increasing because of an upward trend in precipitation. By following the numerical examples in the paper and looking up test statistics in referenced sources, the reader can easily apply these methods to other situations.     

某炼铁厂热风炉系统检测鉴定及加固

通过对某炼铁厂热风炉系统的检测鉴定，对检测数据进行分析，最后根据测试，计算，分析结果，对热风炉系统采取相应加固处理措施，保证热风炉整个系统的安全生产。     

A comparative analysis between temperature and pressure measurements for early detection of runaway initiation

In this work, we have analysed the use of pressure instead of temperature measurements for the early warning detection of runaway initiation. This is possible due to the fact that our runaway criterion, i.e. div>0, does not depend specifically on which state space variable we are using for divergence calculation. A series of runaway experiments, carried out in a 250 l pilot-scale reactor, has been used to compare the results. In accordance with previous analysis, we show that by using temperature, early detection of runaway initiation is achieved. Analogously to temperature, pressure may be also used for runaway detection. By comparing the different types of reactive systems analysed (vapour and gassy), it can be observed that temperature works better, in terms of earlier detection, than pressure but the differences are more pronounced for vapour than for gassy systems.     

NEW ZEALAND'S NATIONAL WATER QUALITY MONITORING NETWORK - DESIGN AND FIRST YEAR'S OPERATION1

ABSTRACT: The design and implementation of a national surface water quality monitoring network for New Zealand are described. Some of the lessons learned from the first year of operation are also addressed. Underpinning the design, and specified in advance, are the goal and objectives, the data quality assurance system, and the mechanism for data interpretation and reporting. Because of the difficulties associated with the use of a multitude of different agencies, only one agency is involved in field work and one laboratory undertakes the analysis. Staff training has been given a high priority. The network has been designed to give good trend detectability for regular sampling over a 5–10 year period.