一种全自动提取居民地的方法

对遥感影像中居民地提取的各种方法作了一般性的阐述,并在吸收这些算法优势的同时,提出了一种全自动实现遥感影像中居民地目标的提取方法：先求取非最大值抑制后的梯度图,而后进行高斯模糊处理,以提高影像上居民地区域纹理特征一致性;通过增强大津法和形态学的方法全方位腐蚀与膨胀,完成面状地物居民地的提取;最后进行图象矢量化处理。论述方法的同时,展示和分析了实验结果。     

一种基于小波域的分形遥感影像压缩方法

解决大数据量遥感影像传输和存储问题的方案之一,是对遥感影像进行高保真快速压缩。通过对遥感影像进行小波变换产生子带频域之间的相似性,再利用分形方法对具有相似性的相邻高频子带进行编码,由高一级的子影像构造低一级的子影像,然后采用分形迭代解码和小坡变换的逆变换重建影像。实验结果表明,该方法在一定程度上能够提高压缩比,在压缩时间上比传统的分形方法有明显的缩短,并且压缩后影像信息的损失量较少,影像恢复的质量较好。     

遥感地质分析在油气靶区建立中的应用

在对松潘-阿坝地区遥感图像解译的基础上,基于GIS对解译出的线性体进行特征分析,并与本区航磁、重力及大地电磁测深资料进行综合分析,从而推断出：根据卫星影像解译出的盆地南部和北部的分界线可能是一条基底断裂,断裂的存在对该地区的油气生成,运移及聚集都可能产生明显的影响。这一推断为在研究区内建立油气勘探区域提供了有力的证据。     

Hierarchical decomposition of variance with applications in environmental mapping based on satellite images

A quadtree-based image segmentation procedure (HQ) is presented to map complex environmental conditions. It applies a hierarchical nested analysis of variance within the framework of multiresolution wavelet approximation. The procedure leads to an optimal solution for determining mapping units based on spatial variability with constraints on the arrangement and shape of the units. Linkages to geostatisiics are pointed out, but the HQ decomposition algorithm does not require any homogeneity criteria. The computer implementation can be parameterized by either the number of required mapping units or the maximum within-unit variance, or it can provide a spectrum of significances of nested ANOVA. The detailed mathematical background and methodology is illustrated by a salt-affected grassland mapping study (Hortobágy, Hungary), where heterogeneous environmental characteristics have been sampled and predicted based on remotely sensed images using these principles.     

基于遥感影像大数据、卷积神经网络的福建省有效致灾雷电分布模型及应用

为了研究福建省有效致灾雷电的分布情况，基于福建省2004—2012年闪电定位数据及雷击人员伤亡数据、福建省L17级谷歌遥感影像瓦片，引入卷积神经网络模型对遥感影像所在区域是否属于人员活动的属性进行建模、训练和预测，得到福建省人员活动属性的格点产品，结合福建省历史雷电数据对有效致灾的雷电分布情况进行了分析，结果表明：①设计的遥感影像+CNN识别模型具有一定的可行性和准确率，通过显著性水平为0.01的假设检验；②福建省有63.55%的格点为无人员活动区域；③平均有45.36%的闪电落在无人员活动的区域，因地制宜地对其他致灾闪电进行预警是提高应急减灾服务效果的可行途径；④有效致灾雷电密度与历史雷击人员伤亡数据的相关性远大于常规雷电密度与历史雷击人员伤亡数据的相关性，有效致灾雷电分布在表征雷电灾害上比常规雷电分布更具有指示意义。     

基于FPN Res-Unet的高分辨率遥感影像建筑物变化检测

针对城市土地资源变化检测工作繁杂、工作量大、自动化程度低等问题，本文提出一种基于深度学习模型的高分辨率遥感影像建筑物变化检测方法，将语义分割的思想引入到遥感变化检测。基于残差结构特征较卷积层提取性能更优和特征金字塔网络多尺度预测的特点，将残差结构和特征金字塔网络融合到Unet模型中，建立FPN Res-Unet模型。该模型以Unet为基础，引入ResNet18的残差结构作为编码路径特征提取层，在每次卷积后使用边界填充，使得输入图像和输出图像尺寸一致；在解码路径每级上采样过程中，拓展支路径将特征金字塔网络融合到模型的网络主干中，将残差结构、Unet及特征金字塔网络的优点相互融合，增强了Unet的特征提取，弥补了语义分割网络对小目标检测的欠缺；在获取深层语义信息的同时关注细节信息，提高建筑物变化检测精度。实验表明，该方法在所用数据集，准确率、召回率、F₁ 3种指标均达到90%以上。     

GIS技术支持下的招远市金矿区崩塌遥感调查

在总结前人研究成果的基础上,全面系统地分析了诱发招远市金矿区崩塌的自然和人为因数,然后运用遥感技术对金矿区遥感图像进行处理,提取诱发崩塌的条件因子,聘请有经验的专家对各项诱发因子进行诊断分析,以此作为判断条件,在G IS技术的支持下对诱发崩塌的条件因子进行空间分析,预测出招远金矿区发生崩塌的危害程度,为防治崩塌提供科学依据。预测结果表明,招远金矿区崩塌有进一步发展的可能,其中,高危险区有36820 m2,中易发区有50 610 m2,低易发区有67 200 m2,需要采取有效措施加以防治。     

1993～2003年厦门市湿地动态变化及其驱动因素分析

通过对厦门市1993年TM遥感影像和2003年ETM 遥感影像的解译分析和数据统计,从湿地面积动态变化和湿地消长情况等方面,用单一土地利用动态度等指标对1993~2003年期间厦门市的湿地变化进行了研究。结果表明,1993~2003年期间厦门市天然湿地面积减少了2103hm2,其中有954hm2(滩涂864hm2)转变为非湿地,有1149hm2转变为人工湿地;而人工湿地面积增加了2775hm2,其中养殖场面积从1993年的5778hm2迅猛增加到2003年的9768hm2,年均增长率为6.91%,水库与水塘和盐田则分别减少了961hm2和254hm2。有27.74%、11.41%、9.33%和3.63%的养殖场面积分别是由非湿地、滩涂、盐田和水库与水塘转变而来的;有54.37%的盐田转变成了养殖场,有33.24%、30.24%的红树林转变为非湿地、养殖场,有10.35%、8.64%的滩涂转变为养殖场、非湿地。对社会经济驱动因素的分析表明,经济发展、人口增长和农业生产比较利益因素极大影响着厦门市湿地的动态变化。     

Towards Formulation of a Space-borne System for Early Warning of Floods: Can Cost-Effectiveness Outweigh Prediction Uncertainty?

The three most important components necessary for functioning of an operational flood warning system are: (1) a rainfall measuring system; (2) a soil moisture updating system; and, (3) a surface discharge measuring system. Although surface based networks for these systems can be largely inadequate in many parts of the world, this inadequacy particularly affects the tropics, which are most vulnerable to flooding hazards. Furthermore, the tropical regions comprise developing countries lacking the financial resources for such surface-based monitoring. The heritage of research conducted on evaluating the potential for measuring discharge from space has now morphed into an agenda for a mission dedicated to space-based surface discharge measurements. This mission juxtaposed with two other upcoming space-based missions: (1) for rainfall measurement (Global Precipitation Measurement, GPM), and (2) soil moisture measurement (Hydrosphere State, HYDROS), bears promise for designing a fully space-borne system for early warning of floods. Such a system, if operational, stands to offer tremendous socio-economic benefit to many flood-prone developing nations of the tropical world. However, there are two competing aspects that need careful assessment to justify the viability of such a system: (1) cost-effectiveness due to surface data scarcity; and (2) flood prediction uncertainty due to uncertainty in the remote sensing measurements. This paper presents the flood hazard mitigation opportunities offered by the assimilation of the three proposed space missions within the context of these two competing aspects. The discussion is cast from the perspective of current understanding of the prediction uncertainties associated with space-based flood prediction. A conceptual framework for a fully space-borne system for early-warning of floods is proposed. The need for retrospective validation of such a system on historical data comprising floods and its associated socio-economic impact is stressed. This proposal for a fully space-borne system, if pursued through wide interdisciplinary effort as recommended herein, promises to enhance the utility of the three space missions more than what their individual agenda can be expected to offer.     

长江口水域多光谱遥感水深反演模型研究

利用Landsat-7 ETM 遥感影像反射率和实测水深值之间的相关性可以探测水深。该文介绍单波段、双波段比值和多波段3种线性回归模型以及动量BP人工神经网络水深反演模型。选择长江口北港河道上段作为研究区,利用上述模型,分两种情况进行水深反演:一是以河道全部历史样本建模;二是将河道按自然水深划分为浅水区和深水区分别建模。结果表明:神经网络模型预测精度高于线性回归模型;水深分区后线性回归和神经网络模型预测误差均有所减小。