【作者】 韩阳；

【导师】 赵云升； 王野乔；

【作者基本信息】 东北师范大学 ， 地图学和地理信息系统， 2011， 博士

【副题名】利用多角度偏振高光谱信息与MODIS影像数据

【摘要】 长白山地区大部分是森林生态系统,是我国东北地区和东北亚地区最重要的生态屏障,也是我国北方生态环境最好的区域之一。在长白山森林生态系统中,森林土壤是整个系统的重要基础,而土壤含水量是土壤中的一个重要组成部分,对陆地与大气间的热量平衡、陆地表面大气环流和土壤温度均产生显著的影响,同时也是评价土壤资源优劣的主要特征之一。运用遥感技术的宏观、准确、动态等显著优势来研究长白山地区森林土壤,有着常规监测方法不可替代的优点,尤其是利用多角度、多时相、高光谱、偏振信息来获取土壤在2π空间内的三维光谱特征,是当今遥感技术的重要研究发展方向。本论文从微观角度和宏观角度出发,对长白山森林土壤同时进行研究与分析。遥感技术的基础是物体对电磁波所呈现的固有特性(包括波谱强度和偏振特性以及物体本身的集合特性特征)的描述。从微观角度出发,本文对长白山地区森林土壤进行了多角度偏振高光谱的光谱测量,充分考虑光线天顶角、光线方位角、探测天顶角、探测方位角以及波长的变化对土壤特征的影响,通过偏振反射特征获得与其对应的二向性反射信息、土壤含水量信息等,对长白山森林土壤建立土壤含水量与光谱信息的数学模型,同时把多角度遥感、偏振遥感以及高光谱遥感结合起来,为定量遥感的发展提供有力地支持。从宏观角度出发,本文利用MODIS产品数据中的NDVI与LST,建立TVDI三角形,构成三角形的“干湿边”,反演长白山地区大范围的土壤湿度,与实测数据相对比分析。论文主要工作内容与结论如下:(1)长白山森林土壤不同含水量条件下的多角度偏振高光谱特征通过对比分析,发现干燥的森林土壤没有明显的偏振反射特性;入射角度一定时,森林土壤的偏振反射的总体趋势是随着探测角的减小而略有减小,波形变化不大,而且不会因探测角度的改变而产生不同的偏振反射特征。在不同含水量的状况下,森林土壤的偏振反射曲线发生变化,当探测角与入射角度相同时,在180°方位角时的偏振反射比时最大的;不同偏振状态下,偏振角45°(即二向性反射比)时的偏振反射比位于偏振角度为0°和90°的偏振反射曲线中间;在入射角、探测角、方位角和偏振状态一致情况下,不同含水量土壤光谱曲线产生差异,随着含水量的增大,其多角度偏振高光谱反射比减小,达到某含水量后,随着含水量的增大其反射比也随之增大。(2)建立多角度偏振高光谱与土壤含水量之间的回归模型经过对长白山地区森林土壤众多曲线的研究,发现森林土壤多角度偏振高光谱具有4大基本特征,其中与水分相关的是1450nm附近的吸收谷和1940nm附近的吸收谷。对这两个水分吸收谷的光谱值、吸光度、微分形式与不同偏振状态下的含水量进行回归建模,得到不同偏振状态下的模型及其相关系数;并构建1450nm附近和1940nm附近的偏振度与含水量之间的线性模型,得到相关系数,并进行统计检验,发现符合统计规律。利用偏振度与含水量之间的线性模型,确定土壤表现为非朗伯体特性的含水量临界值,并计算得之。(3)设计正交试验,进行交互作用分析对不同含水量条件下的森林土壤的多角度偏振高光谱进行正交试验设计与交互作用分析,设计4因素2水平正交试验表格,发现偏振角、探测角、含水量以及方位角均可以对森林土壤的多角度偏振高光谱曲线产生影响,而且这些影响是通过这4个因素之间的交互作用体现的。结果发现,偏振角与含水量的交互作用对森林土壤多角度偏振高光谱曲线影响特别显著,含水量对其曲线也同样影响特别显著,二者均起到了首要作用;探测角与含水量的交互作用对光谱曲线影响显著,而偏振角对其有一定的影响,其它因素对森林土壤多角度偏振高光谱影响不大。(4)利用MODIS产品数据大范围监测长白山地区土壤水分状况运用MODIS产品数据中的NDVI和LST产品,对长白山地区土壤水分进行相关研究,拟合干湿边方程,得到参数,建立TVDI模型反演长白山地区2007年3月~11月的土壤水分状况,同时使用野外采点数据进行长白山地区土壤水分验证。通过对TVDI和土壤相对湿度的线性拟合结果和T检验,发现线性回归方程为显著,说明TVDI能够反映长白山地区土壤含水量状况,根据获取的遥感信息展示长白山地区土壤水分状况的空间分布,并与具有时间序列的MODIS产品数据对比分析TVDI和地表植被覆盖类型的关系。论文的创新之处在于综合微观角度与宏观角度的同时,分析了长白山地区森林土壤水分状况,获得不同含水量条件下的森林土壤光谱曲线与参数,突破了传统研究中仅仅利用单一指标研究和表达土壤水分信息的片面性;同时运用正交试验设计与交互实验分析的方法,展示多角度因素、偏振因素、高光谱因素与含水量因素的交互影响作用,克服传统单一因素分析光谱曲线的弊端,体现了因素间的交互作用对长白山地区森林土壤多角度偏振高光谱特征的影响。更多还原

【Abstract】 Most of the Changbai Mountains area is the forest ecological system, and also the most important ecological barrier in the northeast China and the northeast Asia. It is one of the best ecological environment in the northeast China. The forest soil is the basis of the ecological forest system of the Changbai Mountains, and the moisture is one of the most important parts in the soil. It has the impact on the heat balance between land and atmospheric, land surface atmospheric circulation and the land surface temperature. At the same time it is one of the most important characteristics to evaluate the soil resource. Many advantages in monitoring the forest soil in the Changbai Mountains are revealed by remote sensing technology, especially using the multi-angle information, hyperspectrum and polarized information to obtain the 3-D characteristics of the soil in 2πspace. It is the important direction to research modern remote sensing technology.In this paper, the author researched and analyzed the forest soil in the Changbai Mountains from both micro-perspective and macro-perspective point. The basis of remote sensing technology is to describe the characteristics of the electromagnetic waves, including the spectral intensity, polarized properties and the features of the object itself. From the microscopic point, the author had full consideration to the forest soil spectral in Changbai Mountains, for example, the light zenith angle, azimuth angle, the effects of wavelength and the polarized information. The math-models were built about the forest soil moisture through the polarized reflection characteristics and the bidirectional reflectance information, while take the combination of the multi-angle information, hyperspectral and the polarized information. That could be the strong support for the development of the quantitative remote sensing. From the macro-perspective, the author used the NDVI and Land Surface Temperature (LST) in the MODIS product data and establishes the TVDI triangle. The triangle has the wet and the dry side and we could get the inversion of the soil moisture in Changbai Mountains. And we could get the analysis on the results with the measured data. The content and conclusions of the thesis work are as follows:(1) The characteristics of the multi-angle hyperspectral polarized information of the forest soil in Changbai Mountains with different moisture conditionsIt is found that the dry forest soil has no significant on the reflectance characteristics of polarization through comparative analysis; when the incidence angle is fixed, the trend of the polarized reflection is decreased with the viewing zenith angle decreased and the wave shape has little change. The polarized reflection characteristics would not change with the different viewing zenith angle. The polarized reflection curves changed with the different soil moisture conditions. When the incidence angle equals to the viewing zenith angle, the polarized reflection is the biggest under the relative azimuth 180°; the curve of polarizer 45°is between the angle of 0°and 90°; the same circumstances of the incidence angle, viewing zenith angle, the relative azimuth angle and the polarized angle, the curves have different spectral under different soil moisture. With the moisture increased, the multi-angle hyperspectral polarized reflection is decreased, while reaches some certain moisture content, the reflection is increased with the moisture increased.(2) Regression models between multi-angle hyperspectral polarized information and the soil moistureAfter a large number of curves of the forest soil in Changbai Mountains, it is found that the multi-angle hyperspectral polarized information has 4 major basic characteristics and the associated with the moisture is the reflection valley near 1450nm and 1940nm. The water absorption spectrums in the polarization peaks have the correlation relationship with the moisture under different polarized states. We could also get the models and the correlation coefficient. The models build aroud 1450nm and 1940nm with the degree of polarization and the moisture could go through the statistical test, and we could also find the statistical rules.(3) Making orthogonal design for the interaction analysisThe author made an orthogonal design of the multi-angle hyperspectral polarized information of the forest soil under different moisture, and also had the interaction analysis. It is found that the polarized angle, viewing zenith angle and the relative azimuth angle could have the effects on the multi-angle hyperspectral polarized reflection, and those effects were showed by the interaction. The results showed that the polarized angle and the soil moisture have an especially significant on the curves, the moisture also has the especially significance, and both of the above two played a primary role. The interaction of the viewing zenith angle and the moisture has significant on the curves and polarized angle has particularly significant. All the other factors have little effect on the multi-angle hyperspectral polarized reflection of the forest soil.(4) Monitoring the soil moisture conditions in Changbai Mountais area by the MODIS product dataThe author analyzed the soil moisture in Changbai Mountains using the NDVI and LST information in MODIS product data. We established the TVDI model by the wet and dry functions and the parameters and inverse the soil moisture conditions from March to November in 2007 in Changbai Mountains area. The significant linear regression equation could be found by TVDI and soil relative moisture and T test, which means the TVDI could reflect the soil moisture conditions in Changbai Mountains area. And the author analyzed the surface vegetation types and TVDI by the MODIS product data from the spatial distribution of the soil moisture condition according to remote sensing information.The innovation of the paper is to analysis the soil moisture conditions under both micro level and the macro level and could obtain the forest soil spectral curves and parameters with different water contents, which breaks the traditional study of expression to use only a single index. At the same time the author used the orthogonal experimental design and interaction analysis. That shows the interaction factors of the multi-angle factor, polarized factor, hyperspectral factor and the moisture factor. The paper reflects the interaction of the factors on the multi-angle hyperspectral polarized characteristics of the forest soil in the Changbai Mountains area, which overcomes the disadvantages of traditional single factor analysis.更多还原