山体基面高度的提取方法 ——以台湾岛为例

 山体基面高度的差异影响山体自身对其水热条件的再分配,进而影响山地垂直带谱的结构和分布,是决定垂直带分布高度的重要因子之一。目前,山体基面高度还没有一个准确科学的定义,也缺乏一个有效的数字化、定量化提取方法。本文以台湾岛为例,使用30m分辨率的ASTER GDEM数据,提出了一种提取山体基面高度的方法。首先,以地形特征与水文特征提取方法获得主山脊线与主山谷线,然后,以地形地貌单元自动提取方法获得山体轮廓界线,再依据提取出的主山脊线、山体轮廓界线及主山谷线,划分山体基面高度分区,依据山体基面分布特征确定各分区的基面高度值,将台湾山地划分出6个不同的山体基面高度(0m、150m、 200m、 600m、630m和650m)。该方法为大范围山体基面高度的快速、准确提取,以及山体效应定量化研究提供了重要的技术支撑。     

基于生成式对抗网络的地形特征线提取方法

地形特征线反映了地形表面的起伏形态,其提取对地理分析和研究具有重大意义.近年来,深度学习为图像处理技术提供了巨大的发展潜力.研究提出了基于深度学习模型——pix2pix的地形特征线提取方法,在山地和高原两种地形下进行了实验,并且将该方法应用于多比例尺底图的地形等高线提取研究.实验结果表明,pix2pix模型不仅在地形特征线提取方面有良好的效果,并且在特定的山地地形下可以进行山脊线和山谷线的语义区分.     

基于DEM的一体化山地特征要素提取

山顶点和山脊线等特征地形要素是构成地表地形及其起伏变化的基本框架,对地形在地表的空间分布具有控制作用。基于DEM研究山顶点、山脊线及其空间组合关系,是DEM地表形态特征研究的重要内容,也是衔接从地形特征分析向山峰等地貌学本源语言的途径之一。本文以四川盆地西南缘与青藏高原过渡地带的川西凉山山原为例,基于山峰-山脊线-控制范围一体化构建的算法策略,识别了山峰和山脊线及其等级、主山脊及其范围。结果表明,研究区内有主峰9座,次峰53座,平均高程2540 m;山脊线230条,其中主山脊9条,平均长度60 km;9大山系,近南北走向,平均控制面积1017 km^2。研究用模糊隶属度方法对算法所提取的主峰、主脉进行精度验证,隶属度介于0.98~1.00和0.37~0.57时提取的主峰、主脉基本吻合算法提取的结果。研究采用一体化山地特征要素提取方法,实现了各山地要素间紧密联系、总体结构与区域地貌特征相对吻合的目标;完成了由栅格单元向地理对象的转变;可以应用于协助地貌类型划分,协同区域地理规划等。     

基于SWAT模型和多源DEM数据的流域水系提取精度分析

流域水系是研究水文水资源、地貌演化和生态环境及水土治理等的基础数据,高精度的水系提取对流域研究十分重要。本文以空间分辨率均为30 m的 AW3D30 DSM、SRTM1 DEM和ASTER GDEM2数字高程模型作为基本的地形数据,基于SWAT模型提取犟河流域水系,通过河网“套合差”、水系相对误差、Google Map水文数据及蓝线河网对提取结果进行误差分析与综合评价,探讨河道剖面和地形特征对水系提取精度的影响。结果表明：① 集水面积阈值是决定河网水系提取精度的关键参数,阈值越大,提取的河网密度越小,反之提取的河网密度越大;② 基于河网密度与集水阈值二阶导数的幂函数与直线相切的数学求值方法确定流域最佳集水面积阈值,能避免最佳集水阈值取值的主观性,提取的河网水系与实际河道相符;③ AW3D30 DSM数据提取的流域河网水系与Google Map高分辨率影像的水系偏差最小,且AW3D30 DSM数据提取的水系与蓝线河网的河网“套合差”和水系相对误差值均最低,能真实反映中低山丘陵山区流域水系发育的疏密程度,吻合度最好;④ 多源DEM数据提取结果均显示为河床比降大和横剖面曲线为窄深式的“V”形河谷提取的水系精度高于河床比降小和横剖面曲线为 “碟”形河谷的提取精度;⑤ AW3D30 DSM数据的地形起伏和坡度标准差最大,有利于山区河网水系的提取。因此,基于SWAT模型和AW3D30 DSM数据提取的山区流域水系可最大限度反映流域水系的真实情况,精度最高,此方法和数据源可应用于中低山丘陵山区流域的水系提取研究。     

DEM栅格单元异质性对地形湿度指数提取的影响分析

 地形湿度指数可定量模拟流域内土壤水分的干湿状况,是静态土壤含水量的最常用指标,具有明确的物理意义。但是,由于DEM本身的结构特点,其提取的地形湿度指数具有尺度依赖性。本文主要探讨因DEM水平分辨率不同而导致的DEM栅格单元异质性,对地形湿度指数提取的影响。以厦门市地貌类型比较复杂的西源溪流域为实验区,使用1 ∶1万等高线生成的2.5m和20m分辨率DEM数据,分别提取地形湿度指数并计算栅格单元地形异质性指数,分析DEM栅格单元异质性指数与地形湿度指数之间的关系。研究表明,基于高程标准差、地势起伏度、景观破碎度和多样性的栅格单元异质性指数与地形湿度指数偏差之间均存在显著的负相关性,这4个异质性指数对地形湿度指数差值的对数回归模拟效果良好且显著有效。这对低分辨率DEM提取地形湿度指数的误差纠正,以及描述区域土壤含水量等地形湿度指数的应用研究具有积极意义。     

基于DEM的黄土崾岘提取及其地形特征分析

崾岘是将要被切穿的鞍部,是正负地形矛盾斗争的结果,也是重要的地形控制点。典型的崾岘多位于黄土高原黄土地貌区,又称黄土崾岘,其对识别沟间地与沟谷的斗争程度有一定的指示作用。本文以黄土高原样区为例,基于1:1万DEM（5 m分辨率）和影像分辨率为0.95 m的遥感影像,利用流域边界算法和缓冲区标定,分析窗口选择5×5,实现了崾岘点位的半自动化提取。并对各崾岘点位求取坡度等地形因子,总结崾岘的空间格局和地形特征。结果显示,崾岘多分布在主流域边界和垂直于主沟道的最宽部分,地形控制作用明显。崾岘的坡度、起伏度、切割深度等值均大于鞍部值,同时,高级流域区的崾岘值大于低级流域区的崾岘值,反映出崾岘具有侵蚀程度强、表层完整性低、地表破碎度高的特点。总体而言,崾岘受沟道蚕食度高,从侧面反映了黄土地貌的发育阶段,是黄土地貌发育到中期的标志性产物。     

随机森林方法支持的复杂地形区土地利用/土地覆被分类研究

随机森林方法目前已经成为遥感分类机器学习中一种有效方法,探索基于中等分辨率的Landsat卫星数据与随机森林方法相结合对复杂地形区长时间序列数据的获取及土地利用/土地覆被变化及模拟研究是非常有意义的。本文基于Landsat8OLI卫星多光谱数据,采用随机森林分类方法对青海省湟水流域复杂地形区土地利用类型进行了分类研究。针对复杂地形区域的情况,将研究区进行地理分区,根据每个分区的特点,选择相应的地形特征参数,并通过提取Landsat 8数据的光谱信息与纹理信息构建最优特征集,探索随机森林方法在复杂地形区土地利用分类的适用性。结果表明：使用Landsat8OLI数据进行随机森林分类,能较好地得到湟水流域复杂地形区域的土地利用类型结果;光谱、地形及纹理信息的结合在不同分区的表现结果不同。在脑山区光谱与地形信息结合能使随机森林分类效果最佳,总体精度达到91.33%,Kappa系数为0.886;而在浅山区与川水区综合考虑光谱、地形、纹理信息进行随机森林分类效果最佳,浅山区与川水区总体精度分别达到92.09%和87.85%,Kappa系数分别为0.902和0.859;利用随机森林算法进行优化选择纹理特征组合可以在保证分类精度的同时能够快速地提取土地利用类型信息,为复杂地形区土地利用类型的区分提供了实际可行的方法。     

于DEM光照晕渲模拟的梯田自动提取方法

梯田作为黄土高原最为典型的人工地貌之一,有重要的农业生产和水土保持价值。传统的梯田自动提取仅限于梯田所在区域的范围划定,未能对梯田田坎线实现有效的自动化提取。鉴于此,本文提出了一种基于光照晕渲模拟的梯田快速提取方法。首先,对无人机航测生成的1 m分辨率的数字高程模型（DEM）进行4个方向的光照晕渲模拟,并相加取平均值;然后,通过适当的阈值对均值图像进行二值化,并掩膜掉沟谷等非梯田区域;最后,基于二值化图像自动矢量化得到梯田田坎线,并通过适当的长度阈值进行碎线过滤提高提取精度。本文以陕西省长武县王东沟流域为实验样区进行了实验,结果表明,该方法提取结果的准确率为89.09%,具有较好的提取精度。此外,对该方法涉及到的参数进行了讨论,表明光照模拟的方向角采用2个正交的对称方向对、高度角采用田坎坡度的反正切值、二值化阈值采用t=180-σ的经验公式,可以满足黄土高原的梯田自动快速提取。     

桂林-阳朔地区DEM地形特征与岩性相关性分析及分类研究

地形地貌是岩性解译的重要信息,地形因子作为描述DEM数字曲面几何特征的定量指标参数,可用来定量化表达不同岩性所在地区地形地貌特征。本文以桂林-阳朔地区为研究区,研究地形因子数学、地质意义,建立岩性与地形因子组合间的定量关联,进而实现岩石类型划分。本文基于ASTERGDEM提取坡度、起伏度等12个地形因子,在分析各个地形因子地质意义基础上,通过聚类分析及方差分析的多元统计分析方法,研究各岩性地形因子特性及其关联性,建立研究区岩性之间的定量差异;此外,利用因子分析方法研究岩性分类过程中的主导因素,确定适宜岩性分类方法以实现定量化岩性分类。实验结果表明：不同岩性、不同地形地貌的地形因子（组合）之间具有显著差异,基于因子分析得到的宏观地形复杂度指数（MTI）以及微观曲率指数（MCI）对岩石类型的分类精度达77.36%。研究表明,地形复杂度等地形因子可用于岩性分类,采用因子分析方法可获取反映地形地貌宏观、微观特征的定量指标,且岩性分类效果良好。     

基于GEE的黄河流域植被时空变化及其地形效应研究

黄河流域作为中国东部平原的生态屏障,研讨其植被覆盖的时空变化有助于生态环境治理。本文利用GEE平台,基于Landsat数据通过像元二分模型反演了1990—2020年黄河流域植被覆盖度（FVC）,并通过Theil-Sen Median趋势分析和 Mann-Kendall检验方法剖析FVC的时空变化趋势,挖掘出FVC趋势变化与海拔、坡度、坡向等地形因子之间的响应关系。结果表明：① 黄河流域FVC整体呈现西北低东南高的空间分布趋势,其中低等FVC占整个流域面积的45%,主要集中于西北部干旱半干旱地区;② 流域中部植被覆盖改善明显,占整个流域的57.07%,西北部和东南部退化程度相对较高;③ 植被覆盖受地形效应影响较为显著,在坡度大于40°及高程（-31~637 m）时高等级FVC占比较高,坡度8~18°及高程1852~2414 m范围内植被改善效果相对较好。结果可以为黄河流域生态环境保护及高质量发展提供科学支撑。     

基于垂直带谱的太白山区山地植被遥感信息提取

遥感数据因其全覆盖的优势被广泛应用于山地植被信息的调查和研究。为了实现山区植被类型的高精度提取,本文以太白山区为实验区,结合山地植被的垂直地带性分布规律,利用太白山植被垂直带谱、高分辨率遥感影像（GF1/GF2/ZY3）和1:1万的数字表面模型（Digital Surface Model, DSM）数据,进行了多层次、多尺度的影像分割,构建了具有植被垂直带谱信息的地形约束因子,并据此进行样本选择和面向对象的分类,分类总精度达92.9%,kappa系数达到0.9160。该方法相比于未辅以垂直带谱信息的分类,总精度提高了10%。研究结果表明,分类过程中加入具有垂直带谱信息的地形约束因子,能显著地提高样本选择的效率和准确率,为后续的植被分类提供了精度的保证。通过人机交互的方式,将垂直带谱知识应用到分类中,可以有效地提高山地植被分类的精度。     

Digital Spectra and Analysis of Altitudinal Belts in Tianshan Mountains, China

Based on the framework of the geo-info spectra of montane altitudinal belts, this paper firstly reviews six classification systems for the spectra of mountain altitudinal belts in China and considers that detailed regional study of altitudinal belts is the key for reaching standardization and systemization of mountain altitudinal belts. Only can this further identify and resolve problems with the study of altitudinal belts. The factors forming the spectra of altitudinal belts are analyzed in the Tianshan Mountains of China, and a digital altitudinal belt system is constructed for the northern flank, southern flank, the heartland, and Ili valley in the west. The characteristics of each belt are revealed with a summarization of the pattern of areal differentiation of altitudinal belts.     

Digital extraction of altitudinal belt spectra in the West Kunlun Mountains using SPOT-VGT NDVI and SRTM DEM

In this paper, a digital identification method for the extraction of altitudinal belt spectra of montane natural belts is presented. Acquiring the sequential spectra of digital altitudinal belts in mountains at an acceptable temporal frequency and over a large area requires extensive time and work if traditional methods of field investigation are to be used. Such being the case, often the altitudinal belts of a whole mountain or the belts at a regional scale are represented by single points. However, single points obviously cannot accurately reflect the spatial variety of altitudinal belts. In this context, a digital method was developed to extract the spectra of altitudinal belts from remote sensing data and SRTM DEM in the West Kunlun Mountains. By means of the 1km resolution SPOT-4 vegetation 10-day composite NDVI, the horizontal distribution of altitudinal belts were extracted through supervised classification, with a total classification accuracy of 72.23%. Then, a way of twice-scan was used to realize the automatic transition of horizontal maps to vertical belts. The classification results of remote-sensing data could thus be transformed automatically to sequential spectra of digital altitudinal belts. The upper and lower lines of the altitudinal belts were then extracted by vertical scanning of the belts. Relationships between the altitudinal belts based on the montane natural zones concerning vegetation types and the geomorphological altitudinal belts were also discussed. As a tentative method, the digital extraction method presented here is effective at digitally identifying altitudinal belts, and could be helpful in rapid information extraction over large-scale areas.     

Regional-scale identification of three-dimensional pattern of vegetation landscapes

The altitudinal pattern of vegetation is usually identified by field surveys,however,these can only provide discrete data on a local mountain.Few studies identifying and analyzing the altitudinal vegetation pattern on a regional scale are available.This study selected central Inner Mongolia as the study area,presented a method for extracting vegetation patterns in altitudinal and horizontal directions.The data included a vegetation map at a 1∶1 000 000 scale and a digital elevation model at a 1∶250 000 scale.The three-dimensional vegetation pattern indicated the distribution probability for each vegetation type and the transition zones between different vegetation landscapes.From low to high elevations,there were five vegetation types in the southern mountain flanks,including the montane steppe,broad-leaved forest,coniferous mixed forest,montane dwarf-scrub and sub-alpine shrub-meadow.Correspondingly,only four vegetation types were found in the northern flanks,except for the montane steppe.This study could provide a general model for understanding the complexity and diversity of mountain environment and landscape.     

Magnitude and forming factors of mass elevation effect on Qinghai-Tibet Plateau

Mass elevation effect(MEE) refers to the thermal effect of huge mountains or plateaus, which causes the tendency for temperature-related montane landscape limits to occur at higher elevations in the inner massifs than on their outer margins. MEE has been widely identified in all large mountains, but how it could be measured and what its main forming-factors are still remain open. This paper, supposing that the local mountain base elevation(MBE) is the main factor of MEE, takes the Qinghai-Tibet Plateau(QTP) as the study area, defines MEE as the temperature difference(ΔT) between the inner and outer parts of mountain massifs, identifies the main forming factors, and analyzes their contributions to MEE. A total of 73 mountain bases were identified, ranging from 708 m to 5081 m and increasing from the edges to the central parts of the plateau. Climate data(1981–2010) from 134 meteorological stations were used to acquire ΔT by comparing near-surface air temperature on the main plateau with the free-air temperature at the same altitude and similar latitude outside of the plateau. The ΔT for the warmest month is averagely 6.15℃, over 12℃ at Lhatse and Baxoi. A multivariate linear regression model was developed to simulate MEE based on three variables(latitude, annual mean precipitation and MBE), which are all significantly correlated to ΔT. The model could explain 67.3% of MEE variation, and the contribution rates of three independent variables to MEE are 35.29%, 22.69% and 42.02%, respectively. This confirms that MBE is the main factor of MEE. The intensive MEE of the QTP pushes the 10℃ isotherm of the warmest month mean temperature 1300–2000 m higher in the main plateau than in the outer regions, leading the occurrence of the highest timberline(4900 m) and the highest snowline(6200 m) of the Northern Hemisphere in the southeast and southwest of the plateau, respectively.     

Characterizing the Mass Elevation Effect across the Tibetan Plateau

It is over 110 years since the term Mass Elevation Effect(MEE) was proposed by A. D. Quervain in 1904. The quantitative study of MEE has been explored in the Tibetan Plateau in recent years; however, the spatial distribution of MEE and its impact on the ecological pattern of the plateau are seldom known. In this study, we used a new method to estimate MEE in different regions of the plateau, and, then analyzed the distribution pattern of MEE, and the relationships among MEE, climate, and the altitudinal distribution of timberlines and snowlines in the Plateau. The main results are as follows:(1) The spatial distribution of MEE in the Tibetan Plateau roughly takes on an eccentric ellipse in northwestsoutheast trend. The Chang Tang Plateau and the middle part of the Kunlun Mountains are the core area of MEE, where occurs the highest MEE of above 11℃; and MEE tends to decreases from this core area northwestward, northeastward and southward;(2) The distance away from the core zone of the plateau is also a very important factor for MEE magnitude, because MEE is obviously higher in the interior than in the exterior of the plateau even with similar mountain base elevation(MBE).(3) The impacts of MEE on the altitudinal distribution of timberlines and snowlines are similar, i.e., the higher the MEE, the higher timberlines and snowlines. The highest timberline(4600–4800 m) appears in the lakes and basins north of the Himalayas and in the upper and middle reach valleys of the Yarlung Zangbo River, where the estimated MEE is 10.2822℃–10.6904℃. The highest snowline(6000–6200 m) occurs in the southwest of the Chang Tang Plateau, where the estimated MEE is 11.2059°C–11.5488℃.     

山地垂直带谱数字识别的技术实现和图谱构建

山地垂直带谱是地学信息图谱的一个重要组成部分。"带谱数字识别"就是利用现代数字方法和数据,对客观存在的山地垂直自然带谱进行提取,获取比较完备的山地垂直带连续图谱模式。这是从传统山地垂直带研究走向地学信息图谱研究的重要步骤,有助于使我们对山地垂直谱的认识上升到地学信息图谱的高度。本文探索MATLAB语言快速、准确地实现各种识别算法,结合VB.NET构建的用户操作界面,实现带谱的信息提取、分类集成和多样化表达。山地垂直自然带谱的数字识别分为3种不同的模式:山系单侧、山体单峰和山体多峰;依据各自不同的识别算法,可以得到山地连续的带谱模式。本文全面介绍了数字带谱、带谱数字识别的含义,以及数字识别模型的原理、算法和原型系统等实现过程。     

山地垂直带谱信息系统应用分析与技术改进

山地垂直带信息系统虽实现了山地垂直带谱的首次数字集成,但系统中还存在标识不清、垂直带颜色不易变动、大量带谱显示时可视程度较差等问题。本文以山地垂直带谱空间可视化为研究对象,探讨系统的技术改进,包括山地垂直带谱图形数据的立体标识、翻转页面、颜色可定制、数据轴可操作等功能的实现。山地垂直带信息系统空间可视化功能的进一步完善,不仅可以有效地管理山地垂直带谱数据,而且可方便直观地显示山地垂直带谱数据,有效地集成山地垂直带谱数据并且充分地发掘山地垂直带谱空间变化规律,也为建立界面友好的世界山地垂直带信息系统奠定了基础。     

Tree population dynamics of three altitudinal vegetation communities on Mount Cameroon (1989�C2004)

Changes in permanent sample plots in the lowland,submontane and montane forests on Mount Cameroon(4,095 m above sea level),an active volcano,are described for 15 years from 1989 to 2004.Throughout the study period,the stocking level of trees with a diameter at breast height(DBH) ≥ 10 cm in the three forests were lower than in pan-tropical stands suggesting a significant impact of volcanic and human-related activities on the vegetation communities on the mountain.Annual mortality rates in the submontane and montane forests were consistent with those reported for comparable altitudinal ranges in the Blue Mountains of Jamaica.The annual mortality rate was higher in the lowland forest than other lowland sites included.Divergence between recruitment and mortality rates was large suggesting that the three vegetation communities have not reached their climax.The seven-year difference in half-life of large trees(with a DBH ≥ 50 cm) in the submontane and montane forests suggests an altitudinal effect on turnover of larger trees that in turn contributes to the frequent small stature of high altitude forests.There was little evidence of an altitudinal effect on species turnover and growth rate.This finding supports generalizations about the zero effect of growth on the stature of high altitude trees.Understanding forest dynamics is crucially important in the management of tropical montane environmentsand in this instance particularly so given the recent creation of the Mount Cameroon National Park.