基于LBV变换的遥感影像多步骤分类法研究

面对与日俱增的遥感数据,如何快速准确地进行影像分类成为遥感领域亟待解决的问题。本文在前人的基础上提出了一种基于LBV变换的遥感影像多步骤分类方法。L:地物的总辐射水平,集中反映了裸地的信息;B:地物的可见-红外光辐射平衡,是地面水分状况和水体存在的一个良好指标;V:地物辐射随波段变化的方向和速度向量,能集中反映地面植被状况。多步骤分类法遵循由易到难的原则,能克服类别之间的互相影响,从而提高分类精度。文章结合LBV变换和多步骤分类法的优点对影像进行了分类,结果表明该方法简单易行,且能达到良好的分类精度。     

遥感图像线性影纹理解专家系统设计与实现

线性影纹信息是遥感图像中的一类重要信息,线性影纹信息自动提取是遥感图像智能解译的重要研究领域。笔者采用基于特征对象的专家系统技术来完成线性影纹信息的简单类别提取,设计并实现了线性影纹理解专家系统;阐明了系统结构,解译规则获取、表达,基于消息的不确定性推理机及解释机的设计;最后利用系统做了线性影纹类别提取实验,结果表明基于特征对象的专家系统用于线性影纹类别自动提取是切实可行的,并具有较高的提取精度。     

甘肃两河口地区水系沉积物异常评价

对两河口地区1:5万水系沉积物资料处理,圈定了大量元素异常。应用R型因子分析结果,划分元素组合。从元素组合特征方面分析,评价异常找矿意义,确定了7个找矿远景区,其中1号异常为最佳找矿靶区。     

红枫湖夏季分层期间pCO2分布规律的研究

在夏季分层期间对红枫湖南、北湖湖心的水样进行分层采集,同时测定了分层水样的温度、pH、HCO₃-浓度、溶解氧(DO)、叶绿素a(Chl-a)及铵根离子(NH₄+)、硝酸根离子(NO₃-)、磷酸根离子(PO₄3-)的浓度,水体中CO₂的分压(pCO₂)由化学平衡及亨利定律求得。研究结果表明:光合作用、有机质降解及水体热分层是影响红枫湖夏季pCO₂分布的主要因素。其中,温水层CO₂欠饱和是光合作用吸收CO₂引起的,温跃层中pCO₂的急剧增加是光合产物降解产生CO₂引起的。静水层沉积物附近pCO₂最高并且还有持续增加的趋势,说明沉积物中有机质降解是静水层中CO₂增加的主要原因,夏季湖底水温较高加快了沉积物中有机质的降解。分层现象使pCO₂在水体中的分布差别明显,并且使静水层中CO₂得到积累。此外,夏季红枫湖水体中pCO₂的变化与NH₄+、PO₄3-的变化密不可分,表现为温水层中光合作用消耗NH₄+、PO₄3-,有机质降解过程伴随NH₄+、PO₄3-的释放。     

太湖入湖河道沉积物中生物利用磷和营养水平分析

为了解太湖入湖河道的营养状况,研究了太湖西部河流沉积物生物利用磷的组成与分布。研究结果显示,北部沉积物中营养元素较高,南部较低;沉积物中生物利用磷的含量次序为藻类可利用磷(AAP)>NaHCO3提取磷(OLP)>水溶性磷(WSP)>易解吸磷(RDP),其中AAP是重要的生物利用磷,AAP的比例越高,富营养化程度越高。AAP与营养元素的相关性在不同区域河道有所不同,北部河道与总氮(TN)、总磷(TP)相关性较好,中部和南部河道与沉积物有机质总量(TOM)相关性较好。沉积物的生物利用磷受不同污染源影响较大。对比河道沉积物与湖泊沉积物的特征,发现湖泊沉积物中生物利用磷(BAP)/总磷(TP)、藻类可利用磷(AAP)/总磷(TP)都高于河道沉积物,表明湖泊沉积物中的磷更容易被植物吸收。     

构建基于GIS的房产管理

阐述了基于GIS房产管理的关键是建立统一的数据分类与编码标准,核心是建立房屋的空间地理属性信息、自然属性信息、社会人文属性信息的链接,基础是建立房产地理信息数据库。并提出了建立房产地理信息数据库的方法、设计思路和方式。     

基于免疫粒子群优化算法的影像纹理分类

粒子群优化算法是基于群智能的随机全局优化方法,它源于对鸟群简化社会系统的模拟。为了提高标准粒子群优化算法的收敛性能,将生物免疫系统的记忆能力和多样性引入标准粒子群优化算法,提出一种免疫粒子群优化算法。在提取纹理样本Laws纹理能量模板特征、小波特征等纹理特征的基础上,提出针对分类问题的粒子表达方法和群体寻优策略,实现了基于免疫粒子群算法的纹理分类。实验结果表明,与标准粒子群优化算法相比,免疫粒子群优化算法在获取训练样本类别中心时具有较好的收敛性能,并且基于该算法的影像纹理分类具有较高的分类精度。     

Pelagic records from the Equatorial Ninetyeast Ridge and significant environmental events during the past 3.5 Ma

This paper presents pelagic records of planktic foraminifera, as well as data of stable isotope stratigraphy and carbonate stratigraphy since 3.5 Ma B.P. from site ODP758 in the Ninetyeast Ridge of the Indian Ocean. Based on these data, manifestations and related mechanisms of major tectonic and environmental events such as the rapid uplift of the Himalaya Mountains, “middle Pleistocene climatic transition” and “mid-Brunhes dissolution event” in the region are discussed. According to the analysis and comparison of various indices and changes in terms of foraminifera assemblage, paleotemperature, paleosalinity and themocline from site ODP758, the authors deduce that the paleoclimatic changes might correlate with the mid-Pleistocene transition at 1.4–1.7 Ma B.P. The changes of CaCO₃, mass accumulation rates (MAR) of CaCO₃ and non- CaCO₃ MAR indicate that the loaded terrigenous sediments increased at 1.7 Ma, which is in agreement with the uplift history of the Qinghai-Tibet plateau as shown by the available data. The last two changes coincide with the uplift of the Qinghai-Tibet plateau, hence they are called “Qinghai-Tibet movement” (1.7 Ma), and the “Kunlun-Yellow River movement” (1.2–0.6 Ma). The changes of the CaCO₃ content, coarse fraction (> 150 μm) content and planktonic foraminifera biostratigraphy show that strong dissolution of abyssal CaCO₃ occurred in the study region during 0.5–0.4 Ma. The event was consistent with the “mid-Brunhes dissolution event” in the sedimentary records of the Atlantic Ocean, Pacific Ocean, Indian Ocean and Nansha sea area of the South China Sea. __________ Translated from Geological Bulletin of China, 2007, 26(12): 1627–1632 [译自: 地质通报]     

Automatic classification of volcanic earthquakes by using Multi-Layered neural networks

The application of neural networks as classifiers of seismic events is described with the aim of developing an automatic system for the classification of explosion quakes at the Stromboli volcano. The architecture of the network that we trained to identify four different classes of shocks was a Multi-Layer Perceptron, using the Back Error Propagation algorithm. Five different approaches for representing the information embedded in the seismograms, both in the time and in the frequency domain, were considered, and the results compared. The direct use of the time series of the shocks was not satisfactory. The auto-correlation function worked well, but in some cases it was misleading. A better performance was obtained with a frequency domain representation. Finally, the use of the envelope function did not work well. Combining parameters such as the auto-correlation and envelope functions can improve one source of error, but it may introduce new ones. The performance obtained highlights the importance of the data attributes used for the training of the network. Topologies with eight neurons in a single hidden layer gave, on average, the best results among the considered neural network structures. The overall results provide a large number of events (89% with the best performance) correctly classified, indicating that this automatic technique is reliable, and encouraging further applications in the field of volcanic seismology.     

Trace metals in suspended particulate matter and in sediment trap material from a permanently anoxic fjord — Framvaren, South Norway

Geochemical studies of the trace metal concentrations in suspended particulate matter (SPM) and sediment trap material from a permanently anoxic fjord, Framvaren, South Norway in 1989 and 1993 indicate that extremely high concentrations of zinc (max = 183920 mg/kg), copper (max = 4130 mg/kg), lead (max = 2752 mg/kg), and cadmium (max= 8.1 mg/kg) sometimes (1993) occur in the SPM collected in the anoxic water layer. The highest concentrations of Zn occur just below the redoxcline at 22 m water depth (in 1993), and copper, lead and cadmium have maximum concentrations between 30 and 80 m depth, where the amount of total SPM is at a minimum (about 0.3 mg/L). On a mass per volume (g/L) basis, the maximum concentrations of Cd, Cu and Fe occur at the interface (21m) and those of Zn occur just below the redoxcline (22 m depth). The SPM and sediment trap data suggest that the metals are precipitated as sulfide minerals in the anoxic water. The presence of particulate sulfides was confirmed by SEM studies that show the occurrence of discrete metal (Cu, Fe, Pb, and Zn) sulfide particles in size from 10–20 m as well as framboidal pyrites (1–5 m in size). Higher levels of metal sulfides at intermediate depths rather than in the deep water of Framvaren (> 100 m), may be due to input of trace metals by water exchange over the sill in the upper part of the water column. In the deep water, less metal sulfide precipitation takes place due to depletion of trace metals, and the dilution of particulate metal concentrations by organic matter and by the chemogenic formation of calcite.