基于随机森林模型的太湖水生植被遥感信息提取

水生植被作为太湖湿地的重要组分,其数量和范围变化影响着湖泊生态系统的平衡,故利用遥感技术对水生植被的空间分布开展研究有助于太湖湿地生态系统的保护.以Landsat 8多光谱遥感影像为主要数据源,利用光谱指数和图像变换方法构建多个特征变量,结合随机森林(RF)模型,提取太湖水生植被的空间分布.结果表明:(1)通过对比分析训练样本特征值的平均值、标准差和变异系数,NDVI、NDWIF、SR等指数更易于区分开敞水域和沉水植被、浮叶植被和挺水植被;(2)当设置1000棵分类树和4个分割节点的随机变量时,RF分类模型的袋外误分率小于6%,误分主要受SR、MNDWI和NDVI等特征变量影响;(3)通过验证分析,基于RF模型获得的2014年7月太湖水生植被覆盖面积约为306.0km2,分类精度为88.56%(Kappa系数为0.88),主要分布在湖体的东部和南部,以沉水和浮叶植被为主,两者占水生植被覆盖总面积的84.9%.     

What do Δ14C changes across the Gerzensee oscillation/GI‐1b event imply for deglacial oscillations?

High‐resolution studies were performed on late‐glacial sediments from a small lake in western Denmark with respect to lithology, geochemistry, stable isotopes, pollen stratigraphy and radiocarbon dating on terrestrial macrofossils. One purpose was to detect the so‐called Gerzensee oscillation, or the GI‐1b event, in the later part of the Aller8;d warm period, and to describe the environmental impact of this short cooling. The other aim was to test the hypothesis that considerable Δ¹⁴C changes occur over this time, which can be related to ocean ventilation/thermohaline circulation changes. We find that the GI‐1b event had a major impact on both terrestrial and limnic ecosystems: large vegetation changes, increased soil erosion and lowered aquatic production. By correlations to events in the GRIP ice‐core and ¹⁴C patterns in the Cariaco basin we also transformed our ¹⁴C dated record into calendar years to calculate Δ¹⁴C values. The ¹⁴C dates show that the GI‐1b event both preceded, and was part of, the 11 400–11 300 ¹⁴C yr BP radiocarbon plateau, and was followed by the 11 000–10 900 ¹⁴C yr BP plateau; thus the later part of the event coincides with a distinct age decline. This delayed age drop (Δ¹⁴C rise) in relation to the hypothetical triggering mechanism behind the event, decreased ocean ventilation, could be explained by redeposited macrofossils at the onset of GI‐1b. This phenomenon, also seen at the onset of Younger Dryas, may also reflect increased soil erosion and redeposition at the start of cold periods. The independent Cariaco Basin record, however, implies that the very end of the cool GI‐1b event is related to a distinct rise in Δ¹⁴C. Likewise, the ¹⁰Be record from GISP2 shows a distinct rise in the middle of the event, precluding decreased solar forcing as the trigger of the climate event, but making it likely that high cosmic ray flux (low solar activity) may be the cause of the rising atmospheric ¹⁴C content. We thus conclude that the Δ¹⁴C changes over the Gerzensee oscillation (GI‐1b), being one of several coolings during the Last Termination, does not seem to be related to ocean ventilation changes. The reason behind this lack of coincidence between rising Δ¹⁴C and a fairly distinct Northern Hemisphere cooling may be due to the fact that the oceanic changes during some of these coolings are too subtle to give an atmospheric ¹⁴C imprint, or that an anti‐phase relationship between the two hemispheres blurs the Δ¹⁴C signal, or, finally, that a partly unknown mechanism may lie behind such coolings. Copyright © 2000 John Wiley & Sons, Ltd.     

基于多源数据的地表覆盖制图与分析——以广东省广州市为例

地表覆盖数据是关于土地利用信息的重要来源,在地理国情监测、气候变化研究、生态环境评估等方面发挥着重要的作用。本文以广东省广州市作为研究区域,利用OSM数据结合FROM-GLC数据和Landsat-8遥感影像数据,对研究区域地表覆盖进行分类并完成地表覆盖制图。实验结果表明,利用OSM数据进行地表覆盖制图结果总体精度较高,具有可用性,但是OSM数据空白区域占比较大,有数据区域个别要素数量较少,在参考选取样本时,样本代表性不够高,部分类型的地表覆盖分类精度较低。     

新疆北山地区侵入岩遥感识别方法

基于ASTER数据、Landsat 8数据及WorldView-2数据,采用波段组合法、最小噪声分离法、主成分分析法以及波段比值法,结合岩石光谱曲线,综合提取了新疆北山地区侵入岩的岩性特征。对WorldView-2进行主成分分析和最小噪声分离,并将R(PCA2)G(PCA1)B(PCA4)、R(MNF2)G(MNF1)B(PCA1)、OLI数据R(B7)G(B6)B(B5)进行假彩色合成,利用OLI数据中的B5/B2、B2/B1突出花岗岩。这些方法解决了复杂岩性的影像色调差异较小、细节不清晰等问题,降低了岩性划分及解译的难度,使各种岩性界线更加明晰。结合前人的区域地质调查成果,对研究区侵入岩岩性进行了系统解译及分析。多源遥感数据的综合利用,有助于更有效地识别复杂岩性并进行更详细的岩性分类,提高遥感岩性识别的正确率。     

遥感异常与地球化学指标间的映射关系模型探索

采用改进的遥感异常提取方法,以甘肃金川地区的LandSat 8遥感影像数据为基础,结合主成份分析方法及低通滤波后期处理,得到空间规律性较强的烃基及铁染异常;进而对测区的土壤地球化学取样分析数据,利用统计学方法寻找不同元素异常(或地球化学指标)的内在关联,并建立土壤地球化学指标比值与遥感异常之间的关系模型.继续构建出基于...     

基于实测光谱模拟Landsat-8 OLI数据估算非光合 植被覆盖度

定量的估算非光合植被覆盖度（Fractional Cover of Non-photosynthetic Vegetation, f_NPV）对草原生态系统碳储存、植被生产力、土壤侵蚀和火灾监测均具有重要的意义。本文以锡林郭勒草原实测高光谱和样方盖度为数据源,利用NPV（Non-Photosynthetic Vegetation）、PV（Photosynthetic Vegetation）、BS（Bare Soil）的平均光谱通过线性光谱混合模型模拟得到混合场景光谱,寻找区分NPV/PV/BS的敏感性波段,然后分别评价不同多光谱指数与f_NPV的相关性。最后利用野外混合场景实验验证光谱指数估算f_NPV的有效性。在此基础上,探讨基于OLI数据的NDVI（Normalized Difference Vegetation Index）-DFI（Dead Fuel Index）特征空间是否满足三元线性混合模型的基本假设。结果表明：短波红外（SWIR）波段是区分NPV/PV/BS的敏感性波段,以此为基础构建的OLI-DFI指数具备有效区分NPV/PV/BS的潜力。在模拟混合场景条件下,OLI-DFI和MODIS-DFI指数均与f_NPV呈显著相关,决定性系数R²分别为0.84和0.94,均方根误差RMSE分别为0.09和0.05,而NDI和NDSVI指数与f_NPV相关性很低。与模拟混合场景相比,在野外混合场景下OLI-DFI和MODIS-DFI指数估算f_NPV的有效性均有一定程度的降低,R²分别为0.65和0.75,RMSE分别为0.14和0.12。基于OLI数据构建的NDVI-DFI特征空间满足三元线性混合模型的基本假设,可有效的估算f_NPV。     

影像融合方法应用于西部地区地物提取优选分析

针对西北地区地物地区实际需求,首先,对Landsat-8 OLI影像数据进行彩色合成生成彩色合成影像,与Landsat-8 OLI全色波段影像分别利用HSV变换、PCA替换法、Brovey变换、Gram-Schmidt、Color Normalized(CN)变换等方法进行融合实验分析;其次,利用融合影像与高分辨率Google影像中在不同波段上进行二次融合实验分析,并采用定量客观评价法对融合结果进行评价,分析筛选适合该地区最优的影像融合方法。     

Landsat 8多光谱数据辅助下的高分影像多尺度分割

为解决高分影像分割的边缘锯齿性明显等问题,本文以黑龙江省伊春市桦皮羌子林场为研究区开展了有无多光谱数据辅助的高分影像分割对比试验。首先,本文设计了多尺度分割算法的相同尺度参数下分割试验,确定了该算法分割GF-2影像时应采用的最佳同质性准则组合参数;然后,基于影像分割对象同质性局部方差变化率反映最优分割尺度的思想,利用ESP2工具找出固定尺度范围内的最优分割尺度范围;最后执行最佳同质性准则组合参数配合下的最优分割尺度范围内各个尺度下的多尺度分割,并采用矢量距离指数、紧密度指数、形状指数对2种分割试验结果进行了评价。结果表明,与GF-2影像独立分割相比,Landsat 8多光谱数据辅助下的GF-2影像分割在矢量距离指数、紧密度指数、形状指数的质量上均有提升,平均提升率分别为8.05%、28.40%、11.76%。     

Evaluating metrics derived from Landsat 8 OLI imagery to map crop cover

Developing techniques are required to generate agricultural land cover maps to monitor agricultural fields. Landsat 8 Operational Land Imager (OLI) offers reflectance data over the visible to shortwave-infrared range. OLI offers several advantages, such as adequate spatial and spectral resolution, and 16 day repeat coverage, furthermore, spectral indices derived from Landsat 8 OLI possess great potential for evaluating the status of vegetation. Additionally, classification algorithms are essential for generating accurate maps. Recently, multi-Grained Cascade Forest, which is also called deep forest, was proposed, and it was shown to give highly competitive performance for classification. However, the ability of this algorithm to generate crop maps with satellite data had not yet been evaluated. In this study, the reflectance at 7 bands and 57 spectral indices calculated from Landsat 8 OLI data were evaluated for its potential for crop type identification.     

Evaluation of ICA and CEM algorithms with Landsat-8/ASTER data for geological mapping in inaccessible regions

Many regions remain poorly studied in terms of geological mapping and mineral exploration in inaccessible regions especially in the Arctic and Antarctica due to harsh conditions and logistic difficulties. Application of specialized image processing techniques is capable of revealing the hidden linear mixing spectra in multispectral and hyperspectral satellite images. In this study, the applications of Independent component analysis (ICA) and Constrained Energy Minimization (CEM) algorithms were evaluated for Landsat-8 and Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) remote sensing data for geological mapping in Morozumi Range and Helliwell Hills areas, Northern Victoria Land (NVL), Antarctica. The results of this investigation demonstrate the capability of the two algorithms in distinguishing pixel and subpixel targets in the multispectral satellite data. The application of the methods for identifying poorly exposed geologic materials and subpixel exposures of alteration minerals has invaluable implications for geological mapping and mineral exploration in inaccessible regions.