Visual interpretation of black and white enlargements of landsat imagery for the preparation of a small scale soil map of a part of karimnagar district,andhra pradesh

Individual bands of black and white photographic enlargements of Landsat imagery were interpreted visually using tone, texture, drainage pattern and morphological expressions of the relief features. The results have been related to the existing soil map prepared by API techniques with limited field check. Interpretation map of band 5 gave maximum number of delineations which have relevance to the soil association boundaries of the API map followed by band 4, and bands 6 and 7 helped in delineating water bodies and hills. A composite map prepared by superimposition of the interpretation maps of all the four bands was compared with the existing API map. 12 categories could be delineated on the Landsat composite map as against 13 of the API map. The overall classification performance achieved was in the order of 41.2, 47.1, 93.3 and 30.4 percent respectively for R, U, V and A landscapes and the combined overall performance for all the landscapes was 53%. The classification performance results could be improved by delineating the landscape units at a higher level such as hills, interfluves and toe slopes instead of hills, inselbergs, foot slopes, interfluves and toe slopes.     

Partial polygon pruning of hydrographic features in automated generalization

This article demonstrates a working method to automatically detect and prune portions of waterbody polygons to support creation of a multi‐scale hydrographic database. Water features are sensitive to scale change, therefore multiple representations are required to maintain visual and geographic logic at smaller scales. Partial pruning of polygonal features – such as long, sinuous reservoir arms, stream channels too narrow at the target scale, and islands that begin to coalesce – entails concurrent management of the length and width of polygonal features as well as integrating pruned polygons with other generalized point and linear hydrographic features to maintain stream network connectivity. The implementation follows data representation standards developed by the US Geological Survey (USGS) for the National Hydrography Dataset (NHD). Portions of polygonal rivers, streams, and canals are automatically characterized for width, length, and connectivity. This article describes an algorithm for automatic detection and subsequent processing, and shows results for a sample of NHD subbasins in different landscape conditions in the US.     

Physiography of the tons basin,Western Himalaya

Landsat RBV imagery on 1:500,000 was interpreted to prepare small scale physiography map of the part of the Tons basin. Aerial photographs on 1:60,000 to 1:80,000 scale were interpreted to prepare medium scale physiography map of the basin. The basin has been sub-divided into three physiographic regions viz. High to very high mountains constituting of glaciated and temperate high mountains; Low to moderately high mountains and very low maountains. The physiographic regions have been further sub-divided based on landforms, slope and dissection index. The major physiographic units are summits, repose slopes, serrated ridges with horn/arates; mountain and valley glaciers, morains, solufluction terraces, intrenched incised river valleys, engrown valleys, alluvial terraces, rocky slopes/cliffs and debris.     

基于GIS和分形的云南省富宁县交通网络研究

应用ArcGIS数字化了云南省富宁县扫描地图,获得了该县各等级公路的矢量数据;以县城为中心,计算出该县公路网的长度-半径分数维为1.409 9,结果显示该县公路网密度分布具有显著的分形特征,从县城向周围衰减;将公路矢量数据栅格化,计算出该县的路网分形维数为1.050 5,低于路网规划的1.585;山地地形是该县路网覆盖度较低的主要成因,基于SRTM DEM数据,得出该县坡度≤6°的土地仅占6.435%,地形崎岖导致公路建设的难度与成本增加;最后应用条带剖面法(swath profile),分析了该县广昆高速公路沿线的地形特征,总体地形自云南高原向广西丘陵降低,起伏大,沿线地貌可划分为西部小起伏中山、中部小起伏低山和东部小起伏低山丘陵三段。     

数学形态学在地图模式识别中的应用

扫描数字化是 GIS中快速获取地理数据的手段之一 ,其关键是对扫描图上的地图要素进行自动识别。文中针对地图模式识别的关键技术之一图像分割进行了研究 ,利用数学形态学方法实现了地图上部分要素的分割     

重庆市全域山系脉络梳理及划定工作研究

重庆属于典型的山地地貌类型,但全域山系脉络情况一直没有被全面而详细的梳理。本文研究利用DEM、遥感影像及其他相关资料,从生态角度出发,由下至上地提取重庆市域范围内山体、山脉,形成全市山系脉络基础本底,对重庆市山系生态保护范围的划定及定量化研究具有重要的理论和实践意义。     

Landscape Dynamics in North East Region of India (Meghalaya State) Using Spatial Decision Tree Model

Abstract

India has figured with two hotspots ‐ the Western Ghats and the Eastern Himalayas ‐ in an identification of 8 ‘hottest’ biodiversity hotspots (Myers et. al. 2000). The Meghalaya state (study area) in North Eastern India lies within the “Indo‐Burma” area, which is one of the 8 ‘hottest’ biodiversity hotspots. Timber extraction, the age‐old practice of shifting cultivation and mining has been a major cause of extensive changes in this landscape. Human induced disturbance differs from natural disturbance especially in extension, severity and frequency. Spatial presentation of landscape dynamics can be used to infer disturbance regimes horizontally. Disturbance regimes are mostly dominated by landuse practices in Meghalaya and these landuse practices are important contributors for overall interpretation of ecological processes operating within the landscape. In this scenario, the development of models to study landscape dynamics using remote sensing and GIS would be of great importance to ecologists. In this paper, we make an attempt to characterize landscape dynamics using a decision tree based approach. The varying impacts of human interventions reflected in three zones at landscape level have been brought out in the present study. The physiographic zones of Meghalaya (viz. Garo hills, Khasi hills and Jaintia hills) manifest different landscape characteristics and present varying degree of degradation status. Garo hills, experiences maximum shifting cultivation and has shown highest dynamism in the study area followed by Jaintia and Khasi hills. Characterization of landscape dynamics is important in the perspective of decision makers and policy makers in order to prioritize conservation strategies, so that urgent and necessary action can be taken.     

用专家系统技术实施居民地自动综合

本文主要论述了利用现有地图数据库数据，运用专家系统中产生式规则表示知识的方法，对居民地要素进行自动综合，生成小比例尺地图上居民地数据的方法。     

Quantifying Four Decades of Changes in Land Use and Land Cover in India’s Kailash Sacred Landscape: Suggested Option for Priority Based Patch Level Future Forest Conservation

Effective quantification of land cover changes remains a challenge in Himalayan hills and mountains, and has a colossal value addition for natural resource management. Here we present a new robust method for classifying land cover vegetation at physiognomic scale along steep elevational gradients from ~?200 to ~?7000 masl in the Kailash Sacred Landscape, Western Himalaya, India along with four decades of land use and land cover changes (1976–2011) using remote sensing techniques coupled with intensive ground surveys. Results show that forest cover loss was minimum ca 7.14% of existing forest in 1976; but, however forest fragmentation is high especially in montane broad-leaved and subtropical needle leaved forests. This change largely impacted the quality of valuable tree species such as Quercus spp. Post 1976, continuous migration forced conversion of high altitude agricultural lands into grasslands and scrublands. Human settlement expansion was high especially in low altitudinal range valleys between 1000 and 2000 masl and has increased 6.76 fold since 1976, leading to high forest fragmentation in spite of reduced agriculture area in the landscape. Our physiognomic level classified land cover map will be a key for forest managers to prioritize conservation zones for protecting this unique forest land.     

数字地貌晕渲的特征增强方法研究

小比例尺DEM制作的晕渲图地形破碎,含有大量的干扰细节,不利于从宏观上把握地形的骨架。为了增强数字地貌晕渲图中的地形特征,本文利用滤波和平面曲率相组合的DEM化简方法来描述山谷和山脊区域。分别针对盆地、平原和山地三种地貌类型,采用90 m、100 m和110 m三种分辨率的DEM数据进行实验,研究发现,这三种地貌类型最佳的山脊夸张系数都为1.3,相应的山谷深化系数分别为0.4、0.6和0.8;而三种分辨率情况下山地的最佳山谷深化系数都为0.8,且山脊的夸张系数都为1.3。     

基于航片颜色分割的土地利用更新调查技术

本文对基于航片的土地利用更新调查作业流程进行了分析,对其中的地物提取、地图纠正技术进行了探讨,提出了基于神经元网络和启发式学习规则的彩色地图分层的方法,用以提取彩色地形图中的特征地物以及进行颜色分割。还构造了地图自动纠正算法,它自动准确地提取图内的格网,以进行纠正。这些算法提高了土地利用更新作业效率。     

基于分形理论的水系要素制图综合研究

在研究水系要素具有分形现象的基础上，提出了水系要素的分维数的确定方法，并利用水系的分数维规律进行地图制图综合，取到了较好的结果。     

喜马拉雅山地区冰湖信息的遥感自动化提取

在“全域—局部”分步迭代水体信息提取方法的基础上, 通过对水体信息提取指标——水体指数的物理特性的分析实现了算法中全域阈值的自动选择与局部阈值的自适应调整, 并结合DEM 生成的山体坡度和阴影信息,减少局部迭代过程中对其他地表特征与水体信息的误判。在此基础上, 建立一种适合于高山地区冰川湖泊的自动化提取方案。试验采用Landsat 数据对喜马拉雅山地区的冰川湖泊进行信息提取, 结果表明该方法能够快速准确地完 成大区域范围内的冰川湖泊制图, 并能最大程度地消除高山地区湖泊水体识别中冰川和山体阴影的影响。     

Terrain analysis of Panna District,Bundelkhand, Madhya Pradesh,using LANDSAT imagery

Visual interpretation of LANDSAT imagery in bands 6 and 7 covering Panna district, Bundelkhand, Madhya Pradesh was carried out for terrain analysis and present land use. The physiographic units viz. hills and hill ranges, escarpments, ridges, table-land, pediment, piedmont plain and sub-units of second order were identified and delineated using inter pretation elements. Area of each unit and sub-unit is calculated and presented in the table showing extent, tone, texture, shape and land use. Land use in relation to physiographic units is also described in the paper and shown in the form of small scale map. Visual interpretation technique helps in broad terrain analysis on scale 1:1 million. This information could be used for making a rapid resource inventory.     

基于多尺度特征融合和支持向量机的高分辨率遥感影像分类

相对传统的中低分辨率遥感数据而言,高空间分辨率遥感影像同一地物内部丰富的细节得到表征,空间信息更加丰富,地物的尺寸、形状以及相邻地物的关系得到更好的反映,但其光谱统计特性不如中低分辨率影像稳定,类内光谱差异较大,而传统分类方法仅依据像元的光谱值,因此在高分辨率影像分类中,传统方法往往不能获得好的结果。在此背景下,提出了一种多尺度空间特征融合的分类方法,旨在利用不同尺度的空间邻域特征弥补传统方法的不足。其基本过程是：首先针对不同尺度特点,用小波变换压缩空间邻域特征,并结合支持向量机得到不同尺度下的分类结果,然后根据尺度选择因子为每个像元选择最佳的类别。文中QuickBird和IKONOS影像实验证明该算法能有效提高高分辨率影像解译的精度。     

自适应圆形模板及显著图的高分辨遥感图像道路提取

针对现有基于圆形模板匹配算法需要人工设定模板尺寸的问题,提出一种自适应圆形模板的高分辨遥感图像道路提取算法。首先引入一种改进的局部形态学梯度图,基于形态学梯度图设计了一种自动生成模板尺寸的算法;接着借助改进的道路显著图,按迭代内插的方式搜索起始点与终止点之间其他道路点,搜索过程综合利用了显著图信息和几何夹角信息,使得算法具有更好的识别效果。试验结果表明,本文算法可适用于多种不同条件的遥感图像,能更加有效地从高分辨遥感图像中提取道路信息。     

图形的形态变换和地图代数凸壳算法

讨论并定义了若干关于形态的基本概念,给出了加壳、蜕皮、粘连变换的定义及其实现途径,论证了加壳、蜕皮变换及粘连变换的若干形态特性。联系凸壳概念,论述了运用形态变换实现图形凸壳的原理,并给出了它的通用算法。结合实际算例,阐述了算法的具体步骤和分析结果,并讨论了算法实施中的几个具体理论问题。     

Application of space borne cartographic camera for thematic mapping

A remote sensing experiment ‘TERRA’ was conducted during the first Indo-Soviet joint manned space flight. Under this experiment aerial photographs of the Indian terrain were taken using fixed cameras like MKF-6 and KATE-140 and hand held cameras like Hasselblad 500 EL/M and Practika B-200. In order to find out the positional accuracy of features identifiable on TERRA photographs, an experiment was carried out photogrammetrically using a stereopair of KATE-140 photographs. Analog as well as analytical stereo-plotter were used. A simple technique of comparing the known ground coordinates of various well defined ground features identifiable both on map as well as on stereo model, with the coordinate calculated from the stereopair observation was adopted. Due to the non availability of the actual coordinates of ground details/features and limitations of identifications of such ground details, used for observation in the present study, the results obtained have indicated that the residual errors are random in nature and not systematic. The results showed that such space borne aerial photographs like KATE-140, can be used for obtaining desired cartographic accuracy for the preparation of topographical as well as other thematic maps on scale 1:250,000 and smaller.     

A new landscape metric for the identification of terraced sites: The Slope Local Length of Auto-Correlation (SLLAC)

This work presents the potential for high-resolution remote sensing data (LiDAR digital terrain models) to determine the spatial heterogeneity of terraced landscapes. The study objective is achieved through the identification of a new parameter that distinguishes this unique landscape form from more natural land formations. The morphological indicator proposed is called the Slope Local Length of Auto-Correlation (SLLAC), and it is derived from the local analysis of slope self-similarity. The SLACC is obtained over two steps: (i) calculating the correlation between a slope patch and a defined surrounding area and (ii) identifying the characteristic length of correlation for each neighbourhood. The SLLAC map texture can be measured using a surface metrology metric called the second derivative of peaks, or Spc. For the present study, we tested the algorithm for two types of landscapes: a Mediterranean and an Alpine one. The research method involved an examination of both real LiDAR DTMs and simulated ones, in which it was possible to control terrace shapes and the percentage of area covered by terraces. The results indicate that SLLAC maps exhibit a random aspect for natural surfaces. In contrast, terraced landscapes demonstrate a higher degree of order, and this behaviour is independent of the morphological context and terracing system. The outcomes of this work also prove that Spc values decrease as the area of terraced surfaces increases within the investigated region: the Spc for terraced areas is significantly different from the Spc of a natural landscape. In areas of smooth natural morphology, the Spc identifies terraced areas with a 20% minimum height range covered in terraces. In contrast, in areas of steep morphologies and vertical cliffs, the algorithm performs well when terraces cover at least 50% of the investigated surface. Given the increasing importance of terraced landscapes, the proposed procedure offers a significant and promising tool for the exploration of spatial heterogeneity in terraced sites.     

Geology and geomorphology of Cumbum valley and Varushanadu hills,Madural district,Tamil Nadu through remote sensing

The Varushanad hills forms the eastern offshoot of the Western Ghats and the Cumbum valley divides the Varushnad hills from the Western Ghats. Investigations were carried out in about 1700 sq Km comprising the Cumbum valley and the Varushanad hills of the Western Ghats situated in Periakulam Srivalliputhur and Usilampatti taluks of Tamil Nadu on the geological and geomorphological aspects. Remote sensing techniques were used and limited field checks were made. Visual interpretation of aerial photographs on 1:60,000 scale was carried out for identifying the geomorphological features, drainage pattern and lithology. Digital analysis of the Landsat computer compatible tape (CCT) path-row 154–053 was also carried out to delineate major lithological variations. Major lineaments were also mapped from the Landsat imagery and false colour composite. The rock types of the Cumbum valley and the Varushanad hills are mainly charnockites, granite gneiss and pink granites which have been deformed by folds and faults. The various geomorphological units of the area are ridges, valleys, bazada zone and pediment zone of mountain complex. Resources evalution studies on the potential zone of ground water and possible construction materials were also discussed.