基于区域特征的高分辨率遥感影像变化检测研究

传统像素级变化检测往往忽略邻近有意义的整片区域的空间、纹理、结构等信息,对高分辨率遥感影像具有很大的局限性。本文利用面向对象的思想,提出了一种基于区域特征的特定目标变化检测方法。该方法的技术流程包括:数据预处理;同质区域获取;区域特征选择;同名区域搜索;区域特征比较;变化检测精度评价及变化显示。利用提出的方法对伊朗2003年地震前后的巴姆古城标志性建筑进行检测,总体精度达到89.73%。     

Computerized database on salt affected soils in western and central India using GIS

Salt affected soils occupy significant areas in western and central India manifested by the arid and semiarid climate, sandy/clayey soil texture, absence of natural drainage, and inadequate infrastructure and irrigation development. These soils are productive following reclamation and appropriate management. The National Remote Sensing Agency, Hyderabad (India) published state-wise maps of salt affected soils in India on 1:250,000 scale using a legend that includes physiography, soil characteristics, and the aerial extent of the mapping units. In the analogue form, voluminous data contained in such maps were difficult to handle by users of varied backgrounds. An attempt was made to prepare a computerized database of salt affected soils for easy access, retrieval, and manipulation of spatial and attribute data useful for management of salt affected soils. The salt affected soils maps were prepared, for Rajasthan, Gujarat, Madhya Pradesh, and Maharashtra states, overlaying digitized layers of SAS polygons and the Survey of India basemap using the ILWIS (Integrated Land and Water Information System) software. GIS was used to prepare a composite (master) database of western and central India that showed the extent and distribution of salt affected soils. A relational database was prepared combining the digitized polygons with soil characteristics such as nature and degree of salinity (presence of higher concentration of neutral salts and neutral soil reaction), sodicity (presence of higher concentration of basic salts and alkaline reaction) and ground coverage. The regional and zonal databases of salt affected soils were prepared at a suitable scale overlaying agro-climatic regions agro-climatic zones. Spatial relation of salt affected soils with physiography, climate, geology, and agro-eco-sub-regions were evaluated employing map calculations in GIS. Saline soils were prevalent in Gujarat, and Rajasthan while sodic soils were dominant in Maharashtra and Madhya Pradesh. These were distributed primarily in the arid (B) plain of Rajasthan, alluvial (A) and coastal (D) plains of Gujarat, and peninsular plain (F) of Maharashtra and Madhya Pradesh. It occupied 2,596,942 ha (78%) in the western (Rajasthan and Gujarat) and 733,608 ha (22%) in the central (Madhya Pradesh and Maharashtra) regions. The SAS occupied 3.3 million ha in the western and central region constituting 50% of the total salt affected soils in India. The saline and sodic soils occupied 2,069,285 ha (62%) and 1,261,266 ha (38%), respectively.     

Generation and study of satellite gravity over Gujarat,India and their possible correlation with earthquake occurrences

High-resolution satellite gravity data have been generated and utilized to infer subsurface geological structures in the area of devastating earthquake that struck the Bhuj region in Gujarat on 26 January 2001. Latitudinal gravity profiles have been generated in the Bhuj, Anjar and IBF regions across the epicentres (23.5° N, 69.8° E/M_w 7.0 in 2001; 23.2° N, 70° E/M_w 7.0 in 1956; 24.2° N, 69.2° E/M_w 7.8 in 1819). Substantial differences in gravity anomaly patterns as high as 37 mGal could be observed existing near the epicentre regions. These gravitational differences might have caused due to the plate tectonic processes and due to the changes in densities of different lithospheric zones/sedimentary layers. Temporal variations of the satellite-derived gravity and their probable relations with already occurred major earthquakes in this region have been studied. Hence we conclude that drastic changes in gravity anomalies can be considered as a precursor for occurrences of substantially large earthquakes.     

资源三号数据在省域主体功能区监测方面的应用

当前国家以及各地区主体功能区区域规划已经完成,如何对规划后的主体功能区进行监测,是当今主体功能区监测面临的重大课题。资源三号卫星数据为主体功能区监测提供新的数据源。本文以资源三号卫星数据为数据源,开展主体功能区监测应用,并建立主体功能区监测平台,用于主体功能区监测与地物提取。     

A progressive line simplification algorithm

An application of line simplification considering spatial knowledge is described. A method for identifying potential conflict regions, in order to avoid the self-intersection of generalized, lines, is also discussed. Further-more, a new progressive line simplification algorithm is presented. From the view point of spatial cognition, a spatial hierarchical structure is proposed, and its application to construction of spatial knowledge related to a line is explained.     

基于语义规则的高分辨率影像道路交叉口提取方法

为了准确提取遥感影像上道路交叉口目标,提出一种利用特征语义规则从高分辨率影像上提取道路交叉口的方法。该算法构建交叉口模型时将其视为由同质区域像素集合及区域轮廓边界构成的面对象,提取过程分为两步:1)利用辐射、纹理特征语义匹配提取交叉口候选区域;2)通过几何特征语义匹配筛选候选区域、识别交叉口属性。利用多源遥感影像对算法正确性及合理性进行验证,结果表明:算法能准确、完整地提取道路交叉口,可为影像道路网构建提供辅助信息。     

基于夜间灯光数据的南京城镇用地提取

针对多时相夜间灯光影像无法直接对比的问题,该文提出了一种基于不变目标进行不同年份不同卫星夜间灯光数据的校正方法。在系统校正多时相夜间灯光数据的基础上,应用支持向量机分类算法提取城镇用地信息,并分析1992—2013年南京城镇扩张动态变化规律。结果表明:该数据校正方法可以有效减少年际夜间灯光影像之间的异常差异,提高不同年份数据间的连续性和可比较性;支持向量机分类算法提取的4个样区城镇用地信息总体精度和Kappa系数平均值分别为88.35%和0.56,能够准确反映区域城镇发展的实际情况;1992—2013年南京城镇经历先缓慢后快速的扩张过程,主城区在1992年城镇基础上往四周扩张,并沿长江及南北交通走廊发展。     

基于二维图像特征的点云配准方法

为了提高低覆盖率点云的配准精度和收敛速度,提出了一种基于二维图像特征的点云配准方法。首先采用基于区域层次的点云配准算法实现粗配准;然后将三维点云转换成二维图像,再采用SURF算法提取二维图像的特征,并求解其匹配像素点对;最后根据二维匹配点获取相应的三维点云相关点,并计算刚体变换,由此实现点云的快速精确配准。试验结果表明,与迭代最近点（ICP）算法相比,该点云配准方法的配准精度和耗时分别提高了约20%和60%,是一种快速、高精度的点云配准算法。     

六道湾煤矿塌陷西区地面沉降监测设计方法探讨

介绍乌鲁木齐市六道湾煤矿塌陷区综合治理过程中地面沉降观测站、观测线的设计方法,所需观测内容,观测方法以及所使用观测仪器的精度。在观测站、观测线设计时综合考虑老采空区变形的影响,并针对该区两个不同时段开采变形观测站布设方法分别进行了设计。其中所介绍的设计方法对其它地区类似煤矿塌陷区综合治理有一定意义。     

首都圈地区现今地壳运动的GPS观测与构造活动模拟

利用首都圈地区1992~2004年间280多个GPS点位复测资料,获得了该地区高精度的地壳形变速度场,相对欧亚板块的水平速度场由北向南由4 mm/a逐渐增加到9 mm/a;采用边界元(BEM)理论及位移不连续的多边形元素构建断层的三维模型,并利用均匀介质且各向同性的半无限线弹性空间模型反演计算了断层滑移量,模拟了GPS点位的位移场,在此基础上获得了主要活动断层和首都圈地区的地表、地表下15km和25 km剖面的静态位移场、应力及应变场。结果表明,GPS点位的模拟位移场与观测值在南北向和东西向的误差分别为0.68 mm及0.77 mm,各个主要活动断层在1992~2004年间的构造活动较弱,随着深度的增加,剖面的应变逐渐增加。