滨海湿地信息提取方法比较研究

以江苏省典型滨海湿地为研究对象,利用2005年5月26日的Landsat7 ETM+图像数据,在湿地特征及其遥感图像表征分析的基础上,逐步提高湿地信息的提取精度,通过对多光谱遥感图像特征向量的分析,总结出一些湿地信息提取的规则和方法。在滨海湿地光谱特征分析的基础上,对研究区的图像进行了非监督分类,利用湿地的光谱相应特征、纹理特征、主成分变换、归一化差异水体指数等特征和相应的知识规则,得到用于优化分类的知识规则,采用分层分类的方法对非监督分类的结果进行了优化,从而使提取结果的精度较原来有了很大程度的提高。还利用给予数据挖掘的分层分类法进行分类比较,通过建立误差矩阵和对比各种分类方法的分类精度,总结出一套分类精度较高的针对该研究区的湿地信息提取方法。     

基于ETM+遥感数据的滨海湿地信息提取研究

以江苏省典型滨海湿地为研究对象,利用2005年5月份的Landsat7 ETM+图像数据,在湿地特征及其遥感图像表征分析的基础上,逐步提高湿地信息的提取精度,通过对多光谱遥感图像特征向量的分析,总结出一些湿地信息提取的规则和方法。在滨海湿地光谱特征分析的基础上,首先对研究区的图像进行了非监督分类,进而利用湿地的光谱相应特征、纹理特征、主成分变换、归一化差异水体指数等特征和相应的知识规则,得到用于优化分类的知识规则,采用分层分类的方法对非监督分类的结果进行了优化,从而使提取结果的精度较原来有了很大程度的提高。     

多源信息复合的遥感影像水稻田信息提取方法研究

水稻是中国最主要的粮食作物之一。如何更精确更真实的获取水稻种植信息对于中国农业的可持续发展具有重要的意义。以江苏南京江宁区为试验区,融合遥感影像的光谱信息、纹理信息、空间分布特征等辅助性信息进行基于知识规则的水稻田信息提取,并将提取结果与传统的非监督分类和逻辑通道法的提取结果进行了比较。研究表明,基于知识规则的多源信息水稻田提取方法的精度最高。可见,融合多源信息的基于知识规则分类法是提高遥感水稻田提取精度的有效方法。     

GIS支持下的湿地遥感信息高精度分类方法研究

遥感影像高精度自动分类方法的实现是制约遥感数据应用的瓶颈之一。以知识和地理信息系统为支撑,进行湿地遥感影像的分类,并对各项分类方法的精度进行比较评价,从而为湿地遥感的分类方法提供依据。实验结果表明经辐射增强降噪处理后湿地边界更加明晰;而对于处于生长期的湿地影像,经过光谱增强缨帽处理后,明显提高了区分湿地亚类的精度。结合以上两种分类方法的优势,利用GIS技术对二者进行空间处理,取长补短,生成了湿地遥感影像分类图。实验证明基于3S技术的分类方法精度更高,是一种较好的湿地影像自动分类方法。     

GIS支持下的湿地遥感信息高精度分类方法研究

遥感影像高精度自动分类方法的实现是制约遥感数据应用的瓶颈之一。以知识和地理信息系统为支撑,进行湿地遥感影像的分类,并对各项分类方法的精度进行比较评价,从而为湿地遥感的分类方法提供依据。实验结果表明经辐射增强降噪处理后湿地边界更加明晰;而对于处于生长期的湿地影像,经过光谱增强缨帽处理后,明显提高了区分湿地亚类的精度。结合以上两种分类方法的优势,利用GIS技术对二者进行空间处理,取长补短,生成了湿地遥感影像分类图。实验证明基于3S技术的分类方法精度更高,是一种较好的湿地影像自动分类方法。     

基于多尺度分割的遥感影像滨海湿地分类

基于多尺度的高分辨率遥感影像分类方法研究,可以为滨海湿地动态监测、规划保护提供更详尽的湿地分类信息和更快速的数据获取方法,对湿地保护具有重要意义。选取连云港青口河入海口处湿地为研究区,以高分辨率遥感影像WV\|Ⅱ和航空遥感影像为数据源,利用多尺度分割方法将影像分割成不同层次的实体对象;在不同层次,以实体对象为单元,结合光谱、形状、纹理等不同影像特征,进行滨海湿地分类研究,结果表明:利用该方法分类后,研究区各种湿地类型都达到较高精度。基于多尺度分割的影像分类方法能充分利用各种影像特征完成湿地分类,有效地减少了遥感影像中的“椒盐”现象,提高了分类精度;选择适宜的分割尺度和分割参数是基于多尺度分割的遥感影像分类方法提高精度的前提。     

一种多级再分类技术耕地提取方法

鉴于土地利用中耕地类型的遥感光谱特征差异大,以及我国北方农牧交错带中撂荒地、耕地、裸地和草地混淆严重,耕地信息的获取难度大、精度低,提出了利用长时间序列遥感数据,通过多级再分类技术方法(multilevel reclassification,MLRC)提取可耕种区域。首先利用最大似然法对长时间序列的多期遥感数据进行监督分类,提取出耕地区域,之后在初级分类的基础上,通过统计不同区域在多期分类结果中被判定为耕地的次数进而确定可耕作区域的范围。通过对闪电河湿地实验区的研究表明,利用MLRC方法的精度达到了82.56%。     

基于遥感影像的军事阵地动态监测技术研究

针对部队快速机动作战的军事要求,提出基于高分辨率遥感影像的军用阵地动态监测方法。借助面向对象的多尺度分割技术将阵地影像分割为同质对象,以提取各个对象的特征;针对监督分类和非监督分类的弊端,提出通过一定的先验知识制定分类规则的方法对遥感影像进行地物识别,在此基础上定性和定量地输出变化检测结果。实验结果表明:利用基于对象影像分析方法具有较高的识别精度,能够有效监测军事阵地变化。     

基于关联规则的面向对象高分辨率影像分类

以北京市昌平区Geoeye-1高分辨率遥感影像为试验数据,研究了关联规则挖掘和面向对象相结合的地物分类方法。首先探讨了关联分类法的原理,再通过图像分割、特征提取、关联规则挖掘、分类器构建一系列过程实现了基于关联规则的面向对象高分辨率影像分类,最终评估分类精度并与K-近邻法进行了对比。结果表明,该方法具有较好精度,能够在一定程度上摆脱地物分类对于专家知识的依赖。     

基于决策树方法的淡水沼泽湿地信息提取——以三江平原东北部为例

湿地是生态系统的重要组成部分,及时、准确地获得湿地基础信息,对湿地的动态监测、保护与可持续利用及其它领域的研究具有重要意义。以三江平原东北部沼泽湿地为例,利用分类回归树(Classification and Regression Tree,CART)算法从训练样本数据集中挖掘分类规则,集成遥感影像的光谱特征、纹理特征和地学辅助数据建立研究区湿地信息提取的决策树模型。用实测的GPS样本点对分类结果进行精度验证,并与最大似然监督分类方法(Maximum Likelihood Classification,MLC)进行对比。结果表明,基于CART的决策树分类结果的总精度和Kappa系数分别为82.65%和0.7935,分类精度较MLC监督分类方法有明显提高,是内陆淡水沼泽湿地信息提取的有效手段。     

Long-term vegetation changes in four types of wetland in China and USA between 2000 and 2011: observations from MODIS

Wetland vegetation assessment is a critical component of evaluating wetland protective policies and providing useful information for global climate change research. In this study, four types of wetland are chosen to evaluate their long-term changes and potential linkage to anthropogenic (e. g., agricultural activities) and natural influences (e. g., hydrometeorological effects). These wetland types include the Yancheng wetland (coastal wetland), the Chongming wetland (river wetland), the Hongze Lake wetland (lake wetland) and the Everglades wetland (marsh wetland) in the contrasting environments of China and USA. The vegetation coverages of these wetlands were evaluated from 2000 to 2011 by using over 2,200 Moderate Resolution Imaging Spectroradiometer (MODIS) 250 m resolution images. Four vegetation indices (VIs) were compared to evaluate their effectiveness in assessing relative changes: the Normalized Difference Vegetation Index (NDVI), the Floating Algae Index (FAI), the Enhanced Vegetation Index (EVI), and the Visible Atmospherically Resistance Index (VARI). FAI performance was relatively insensitive in terms of both the statistical error and the aerosol effects compared with other VIs and was thus chosen to study the long-term vegetation changes. The results showed that 1) agricultural influence appeared to be relatively minimal compared with hydrometeorological effects in the Everglades wetland and in the core partitions of the Yancheng, Chongming, and Hongze Lake wetlands over the 12-year period, and 2) the entire partitions of the Yancheng, Chongming and Hongze Lake wetlands showed noticeable influences from agricultural activities in addition to natural variability. The cost-effective method that we demonstrate in this study may be extended to other wetlands near the same geographical location and to other satellite platforms.     

基于面向对象—深度学习的闽东南低海拔海岸带地区湿地动态遥感分析

海岸带湿地具有重要的生态价值和经济开发价值,明确其时空变化特征与影响因素对于维持区域生态系统平衡和可持续发展具有重要意义。以Landsat TM/ETM+/OLI影像为基本数据源,综合利用面向对象与深度学习分类方法对1985~2015年闽东南低海拔海岸带地区的湿地信息进行提取,以揭示其时空演变特征与驱动力因素。结果表明：基于面向对象—深度学习分类方法对湿地进行信息提取,整体分类精度可达93%以上,分类结果整体性好;1985~2015年自然湿地面积呈减少趋势,人工湿地面积呈增加趋势,分别减少和增加250.31 km²和251.36 km²;湿地二级类型中,30 a间河口/浅海水域和淤泥质海滩面积减少最大,盐田/水产养殖场面积增加最大;1985~2015年湿地变化类型多样,且2000~2015年较1985~2000年湿地变化更为剧烈;湿地变化是人类活动和自然环境变化共同作用的结果,其中人类活动是影响湿地变化的主要原因。该方法及研究结果可为海岸带湿地监测与保护管理提供技术支持和决策参考。     

基于多时相遥感和GIS技术的湿地识别研究

本以江苏省洪泽湖湿地为例,首先利用秋季成像的TM影像进行分类,采用了混和分类法。然后利用同年度春季成像的TM影像制作湖泊掩膜来纠正分类中湿地类型与高地类型的错分现象。研究发现,该方法能有效地降低分类的误差,提高分类的精度。     

基于知识规则的洪泽湖湿地信息提取及动态变化研究

湿地是自然生态系统的重要组成部分。快速、准确地获取湿地基础信息,对湿地的动态监测、保护和可持续利用具有重要意义。本文利用知识规则分别提取出洪泽湖1988年10月25日和2002年11月9日两时相遥感影像的湿地信息,然后通过叠加分析比较了两个时相的湿地信息变化情况。研究发现,洪泽湖湿地植被在这14年间急剧减少,有待采取有效的措施进行保护。     

基于面向对象分类方法的乌苏里江流域中俄跨境区域湿地景观动态研究

选取乌苏里江中下游干流50km缓冲区为研究区,以1989、2000和2013年的Landsat TM、OLI遥感影像为数据源,利用面向对象的分类方法获取湿地信息;在GIS技术支持下,选取斑块类别面积(CA)、面积所占比例(PLAND)、最大斑块面积指数(LPI)、平均斑块面积(AREA-MN)、斑块数(NP)、斑块密度(PD)等景观格局指数,分析了研究区1989~2013年湿地的景观格局动态特征。结果表明:面向对象的分类方法能够准确提取湿地信息,并且有效地避免"椒盐现象";1989~2013年,该区水田、水库/坑塘、运河/水渠、灌丛沼泽增加和河流、湖泊、草本沼泽、森林沼泽减少;大量草本沼泽转为水田为该区湿地变化主要特征;1989年湿地类型以草本沼泽为主,2013年水田的斑块面积略高于草本沼泽斑块面积;湿地景观变化区主要分布于中国一侧的挠力河流域和穆棱河流域;天然湿地景观破碎化程度逐年加深,人工湿地景观趋于聚集,人为干扰是导致湿地景观格局变化的主要动力;中俄两国境内湿地景观变化程度差异显著,中国境内大量天然湿地转为人工湿地;俄罗斯境内的湿地变化微小,天然湿地保存较好。农田垦殖是导致两国湿地景观破碎程度差异的主要原因。     

Improved wetland remote sensing in Yellowstone National Park using classification trees to combine TM imagery and ancillary environmental data

The U.S. Fish and Wildlife Service uses the term palustrine wetland to describe vegetated wetlands traditionally identified as marsh, bog, fen, swamp, or wet meadow. Landsat TM imagery was combined with image texture and ancillary environmental data to model probabilities of palustrine wetland occurrence in Yellowstone National Park using classification trees. Model training and test locations were identified from National Wetlands Inventory maps, and classification trees were built for seven years spanning a range of annual precipitation. At a coarse level, palustrine wetland was separated from upland. At a finer level, five palustrine wetland types were discriminated: aquatic bed (PAB), emergent (PEM), forested (PFO), scrub-shrub (PSS), and unconsolidated shore (PUS). TM-derived variables alone were relatively accurate at separating wetland from upland, but model error rates dropped incrementally as image texture, DEM-derived terrain variables, and other ancillary GIS layers were added. For classification trees making use of all available predictors, average overall test error rates were 7.8% for palustrine wetland/upland models and 17.0% for palustrine wetland type models, with consistent accuracies across years. However, models were prone to wetland over-prediction. While the predominant PEM class was classified with omission and commission error rates less than 14%, we had difficulty identifying the PAB and PSS classes. Ancillary vegetation information greatly improved PSS classification and moderately improved PFO discrimination. Association with geothermal areas distinguished PUS wetlands. Wetland over-prediction was exacerbated by class imbalance in likely combination with spatial and spectral limitations of the TM sensor. Wetland probability surfaces may be more informative than hard classification, and appear to respond to climate-driven wetland variability. The developed method is portable, relatively easy to implement, and should be applicable in other settings and over larger extents.     

Classification of Vegetation Coverage of Wetland Landscape based on Remote Sensing in the Coastal Area of Jiangsu Province

The Vegetation Coverage Estimation Model (VCEM) was established based on the classification of coastal wetlands by remote sensing technology and the improved dimidiate pixel model as well as the Normalized difference vegetation index (NDVI).The VCEM was used to calculate vegetation coverage (Fc) and grade level for different wetland types in the study area of Yancheng,Jiangsu Province.The relationship between wetland vegetation coverage and their types was analyzed further by overlaying the coverage classification map and wetland landscape map.The results show that the zonal distribution of plant community of the coastal wetland was notable.And the vegetation coverage level varies with different types of wetland.The reed marshes tend to grow at the high vegetation coverage zone,which account for 67.45% of the total area of reed marshes,and the medium vegetation coverage zone account for 23.61%,and the low vegetation coverage zone only account for 8.94%.Most the Suaeda salsa marshes were in the medium vegetation coverage zone,which account for 81.14% of its total area,and 3.36% and 16.06% were distributed at the high vegetation coverage zone and the low vegetation coverage zone respectively.Most of the Spartina alterniflora marshes were distributed at the high vegetation coverage zone,accounting of 83.47% of its total area,and only 15.07% at the medium vegetation coverage zone,1.74% at the low vegetation coverage zone.In addition,most high vegetation coverage areas of wetlands were distributed at the middle location of each vegetation zone,and the Transition zone for each wetland type was usually with lower vegetation coverage level.