四种遥感浅海水深反演算法的比较

详细介绍了单波段线性回归模型、两波段比值线性回归模型、多波段组合线性回归模型、BP神经网络模型等4种光学遥感水深反演算法,然后利用同一地区、同一时期的Worldview-2多光谱遥感影像和实测水深数据,对4种水深反演模型的准确性进行了实验比较。研究表明:多波段组合线性回归模型、BP神经网络模型的水深反演的性能较好,利用多光谱遥感图像数据反演得到的水深值误差较小;而单波段线性回归模型、两波段比值线性回归模型的效果较差。     

人工神经网络方法在夏季降水预报中的应用

在夏季雨型预报中引进了人工神经网络方法。首先,根据雨型与前期（冬季）环流和海温的关系,从前期冬季资料场中找预报因子;然后,用人工神经网络方法对我国夏季的雨型进行模拟预报,以前４０年资料做训练样本,让网络在一定的学习规则下进行学习,最后得到一种分类预报模型。经对１９９２～１９９６年夏季雨型做独立试报,结果与实况基本相符。     

Prediction of groundwater inrush into coal mines from aquifers underlying the coal seams in China: application of vulnerability index method to Zhangcun Coal Mine,China

Groundwater inrush is a geohazard that can significantly impact safe operations of the coal mines in China. Its occurrence is controlled by many factors and processes are often not amenable to mathematical expressions. To evaluate the water inrush risk, Professor Wu and his colleagues have proposed the vulnerability index approach by coupling the artificial neural network (ANN) and geographic information system (GIS). The detailed procedures of using this innovative approach are shown in a case study. Firstly, the powerful spatial data analysis functions of GIS was used to establish the thematic layer of each of the main factors that control the water inrush, and then to choose the training sample on the thematic layer with the ANN-BP Arithmetic. Secondly, the ANN evaluation model of the water inrush was established to determine the threshold value for each risk level with a histogram of the water inrush vulnerability index. As a result, the mine area was divided into four regions with different vulnerability levels and they served as the general guidelines for the mine operations.     

Hydrologic simulation approach for El Niño Southern Oscillation (ENSO)-affected watershed with limited raingauge stations

ABSTRACT

Since the performance of hydrological models relies on numerous factors, the selection of an appropriate modeling approach for hydrological study has always been a crucial issue. The major objective of this research is to demonstrate that data-driven models such as the Adaptive Neuro-Fuzzy Inference system (ANFIS) are more suitable in a region where spatially distributed precipitation datasets are not available. Since precipitation has a teleconnection with the El Niño Southern Oscillation (ENSO) in different parts of the world, the sea surface temperatures (SSTs) and sea level pressures (SLPs) of the equatorial Pacific can be expected to act as surrogates for the precipitation if there are insufficient raingauge stations in the watershed. Moreover, in contrast to conceptual and physically-based models, data driven models can incorporate SST and SLP in their input vectors, and hence additional forcing of SST with precipitation has been experimented with in past studies. Therefore, our second objective is to test whether the additional forcing of SST and SLP will improve the hydrologic simulation. For this, various ANFIS models for the winter season were developed considering 10 raingauge stations situated at various locations in the watershed. Rainfall from each raingauge station was considered in the ANFIS model one at a time with and without SST/SLP. The results show that the performance of the ANFIS model improved with the additional fusion of SST and SLP, especially when a raingauge station from a remote location was considered. However, this improvement was observed when the analysis was primarily focused on the winter season which is a period with a strong ENSO signal.

Editor D. Koutsoyiannis Associate editor L. See     

城镇基准地价评估的人工神经网络模型研究

在我国的城镇基准地价评估中,通常采用回归统计模型。这类模型属确定型数学模型,受主观影响较大,不能全面准确地模拟地价与其影响因素之间复杂的非线性关系,且评估工作量大,不易实现评估自动化。在此探讨了应用人工神经网络技术建立城镇基准地价评估模型的理论、方法和步骤,分析和研究了模型优化方法,并通过实例验证了模型的适用性和可靠性。研究表明,3层结构的基于误差反传算法的BP网络模型,能较好地模拟城镇基准地价与其影响因素之间的复杂关系。     

三层BP神经网络地震灾害人员伤亡预测模型

选择地震发生时刻、震级、震中烈度、建筑物倒塌和严重破坏率、抗震设防水准、人口密度、地震预报等7个评价指标,以20次严重地震灾害为示例（其中,17个作训练样本,3个作验证样本）,建立了三层BP神经网络地震灾害人员伤亡预测模型。基于MATLAB6,5BP神经网络训练,得出的预测结果与各个示例的实际数值比较吻合。验证样本的训练结果表明,该模型适用于地震灾害人员伤亡评估。通过对评价指标的权重计算,确认人口密度、建筑物倒塌与严重破坏率、震中烈度是影响地震灾害人员伤亡的主要因素,地震预报、抗震设防水准、地震发生时刻和震级次之。作为人为可控预测指标,减少人口密度特别是城市人口密度,提高建（构）筑物抗震能力及预测预报水平,对于减少地震灾害人员伤亡起更重要的作用。     

Fusion of VNIR and thermal infrared remote sensing data based on GA-SOFM neural network

The multi-source data fusion methods are rarely involved in VNIR and thermal infrared remote sensing at present. Therefore, the potential advantages of the two kinds of data have not yet been adequately tapped, which results in low calculation precision of parameters related with land surface temperature. A new fusion method is put forward where the characteristics of the high spatial resolution of VNIR (visible and near infrared) data and the high temporal resolution of thermal infrared data are fully explored in this paper. Non-linear fusion is implemented to obtain the land surface temperature in high spatial resolution and the high temporal resolution between the land surface parameters estimated from VNIR data and the thermal infrared data by means of GA-SOFM (genetic algorithms & self-organizing feature maps)-ANN (artificial neural network). Finally, the method is verified by ASTER satellite data. The result shows that the method is simple and convenient and can rapidly capture land surface temperature distribution of higher resolution with high precision.