遥感组合指数与不同分类技术结合提取农业用地方法

多光谱遥感影像因具有丰富的波谱信息,提高了地表覆盖的辨识能力,利用遥感数据高精度自动提取专题信息是目前研究的热点和难点。本文以北京市ASTER影像为例,通过对城市生态环境中土地类型及其光谱特征规律分析,组合归一化差异植被指数、修正归一化差异水体指数和归一化差异建筑指数三种指数,制作组合指数新影像。对组合指数影像采用基于支持向量机的面向对象分类方法进行农业用地信息提取,同时将该方法分别与基于原始影像、组合指数影像的最大似然及支持向量机的分类方法进行对比分析。实验结果表明:组合归一化差异指数影像压缩了数据维数,降低了覆盖地物相关性,易于农业用地信息提取。对组合指数影像采用基于支持向量机的面向对象分类方法精度达95.701%。     

利用WCF进行空间矢量数据服务模型设计

地理信息网络服务越来越普遍,但矢量数据服务在实际应用中却存在很多问题。针对现有的矢量数据服务接口标准(如WFS)过于简单无法满足复杂的实际应用的现状,设计了一个基于WCF数据契约的空间矢量数据服务模型。针对传统Web Service技术对大数据流处理效率低的问题,提出利用WCF信道压缩编码器提高大数据量传输的方法;同时利用WCF内置的安全模式实现了矢量数据的传输安全。     

利用SVM-CRF进行高光谱遥感数据分类

提出了一种改进的随机场模型SVM-CRF,它以支持向量机作为条件随机场的一阶势能项,结合了支持向量机和条件随机场的优点。采用AVIRIS高光谱遥感数据进行实验,对SVM-CRF模型进行了分析,结果表明,在分类精度上SVM-CRF优于支持向量机和传统条件随机场模型。     

A study of partitioning Q vector on background conditions of a torrential rainfall over Shanghai, China on 25 August 2008

A rainfall that occurred during 0200–1400 Beijing Standard Time(BST)25 August 2008 shows the rapid development of a convective system,a short life span,and a record rate of 117.5 mm h-1for Xujiahui station since 1872.To study this torrential rainfall process,the partitioning method of Q vector is developed,in which a moist Q vector is first separated into a dry ageostrophic Q vector(DQ)and a diabatic-heating component.The dry ageostrophic Q vector is further partitioned along isothermal lines in the natural...     

Ricker子波核最小二乘支持向量回归滤波方法的稳健性研究

除了信噪比、有效子波畸变等,稳健性(Robustness)也是度量滤波方法效果的一个重要的物理量,它刻画了滤波系统应对异常点值的能力.一般用影响函数作为评价稳健性的工具.支持向量机方法已较成功地应用于信号与图像的滤波中,尤其Ricker子波核方法更适于地震勘探信号处理.通过考察Ricker子波核最小二乘支持向量回归(L...     

Fault tectonic analysis of Kii peninsula,Southwest Japan: Preliminary approach to Neogene paleostress sequence near the Nankai subduction zone

The southern part of the Outer Zone of Southwest Japan including the Kii peninsula belongs to the tectonic ‘shadow zone’, where fewer conspicuous active faults and less Quaternary sediments develop than in the Nankai subduction zone and Inner Zone of Southwest Japan. In order to study the paleostress sequence of the Kii peninsula, we analyzed fault‐slip data and tension gashes at pilot sites of Early–Middle Miocene forearc sediments and Late Cretaceous accretionary complex. According to the results, six faulting events are reconstructed in sequence: (i) east–west extension (normal faulting); (ii) east–west compression and north–south extension (strike‐slip faulting); (iii) NNW–SSE compression and ENE–WSW extension (strike‐slip faulting); (iv) northeast–southwest compression and northwest–southeast extension (strike‐slip faulting); (v) WNW–ESE compression (strike‐slip or reverse faulting); and (vi) NNE–SSW extension. The north–south to NNW–SSE trending dyke swarm of Middle Miocene age in the Kii peninsula is thought to be related to Event 3, implying that Event 3 was active at least during the Middle Miocene. Because Event 6 is recognized solely at a site, the overall latest faulting event seems to be Event 5. Assuming that the compression results from the motion of the crust or plate, the compression direction of Event 5 is in good accordance with the present‐day WNW crustal velocity vectors of the Kii peninsula. The stress trajectory map of Southeast Korea and Southwest Japan reveals that the current compression directions of the Kii peninsula correspond to the combinatory stress fields of the Himalayan and Philippine Sea tectonic domains.     

基于GA-SVR的渗透系数参数反演方法

渗透系数参数反演的本质是优化问题求解,遗传算法是一种基于自然选择和群体遗传机理的新的全局优化求解方法,可以较好地用于求解诸如渗透系数参数反演等复杂非线性组合优化问题。基于结构风险最小化原理的支持向量机具有逼近复杂非线性系统、较强的学习泛化能力,可以用来计算渗透系数参数反演过程中的测点水头值。实验表明,基于遗传算法-支持向量回归机的地下水渗透系统参数反演拟合效果良好,能大大提升区间搜索效率,避免出现局部最优解,其参数识别精度符合实际应用要求。     

Research on the Distribution and Content of Water Ice in Lunar Pole Regions Using Clementine UVVIS Data

Interest in the Moon started to increase at the beginning of the 21st century,and henceforth,more and more attention has been paid to the content and distribution of water ice in the lunar polar regions.The existence of water or ice in the regolith can apparently change its dielectric features.Therefore,in this article,the Dobson model is adopted and improved according to the Moon's environmental features,to construct the relationship between the volumetric water ice content and the dielectric constant.Ther...     

煤田井温测量与煤层一、二级热害区的圈定

目前国内煤田地质勘探中垂直深度达1200～1500m,部分省区采煤深度也达到了1000m,地层温度都比较高。一般情况下,一级热害区可以通过通风可以达到降温的效果,而二级热害区就必须在井底进行人工制冷,才能确保工人的生命安全。以陶乐南部勘查区测量井温为例,确定了该区恒温带的深度为60m,温度为14.55℃,并将96.0%作为孔底温度校正系数,计算出各简易测温钻井孔底的校正温度及平均地温校正系数;据此计算平均地温梯度和煤层底板温度。依据煤层底板深度与温度作出的井温等值线,圈定煤层一级热害区（地温达到31～37℃）与二级热害区（地温大于37℃）的分布范围并计算其面积：该勘查区9-1煤层底板一级热害区0.46km2;二级热害区6.05km2。     

基于阵列感应测井的支持向量机流体识别方法

基于研究区块的高分辨率阵列感应测井(HDIL)资料,首先分析了利用交会图法进行流体识别的效果,指出了该常规解释方法的局限性。在此基础上,将基于遗传算法优化的支持向量机(GA-SVM)引入到测井解释当中,结合阵列感应测井和常规测井资料,建立了流体识别的非线性模型。利用该模型对45个训练样本进行回判,准确率为100%,对21个预测样本进行预测的准确率为90.48%,相对于常规解释方法,该流体识别方法具有更高的准确率。