| 题名 | 山区HJ CCD影像气溶胶光学厚度与地表反射率反演算法研究 |
| 作者 | 赵志强 |
| 学位类别 | 硕士 |
| 答辩日期 | 2014 |
| 授予单位 | 中国科学院研究生院 |
| 授予地点 | 北京 |
| 导师 | 李爱农 |
| 关键词 | 山区HJ CCD影像 暗目标 气溶胶光学厚度 大气校正 地表邻近效应校正 |
| 其他题名 | The study on retrieval of aerosol optical thickness and surface reflectance from HJ CCD image in typical mountainous areas |
| 学位专业 | 地图学与地理信息系统 |
| 中文摘要 | 无论是遥感应用的实际需要，还是定量化遥感的发展趋势，都要求遥感影像能够提供与地表实际特性相符的地表反射率信息。但是传感器接收到的遥感信息会受到大气光学特性和地表环境辐射特性的共同影响，而导致遥感影像对比度减小，影像质量降低，地表信息提取难度增加。因此，对遥感影像进行大气校正和地表邻近效应校正、获取真实地表反射率是遥感信息定量化过程中不可缺少的一个重要环节。气溶胶对可见光波段的辐射有强烈的散射和吸收作用，是大气校正中最为关键的大气参数，并且气溶胶在大气辐射强迫和大气环境质量评价等方面也拥有举足轻重的地位，气溶胶光学厚度反演一直是遥感定量研究的热点。中国是一个山地大国，山地面积约占全国总面积的70%，然而山区地形起伏巨大、气溶胶时空变化剧烈、气候多变，且存在大量的地形阴影、云以及云下阴影，这些都为山区遥感影像气溶胶光学厚度反演和地表反射率反演增加的困难。HJ CCD影像是我国专门用于环境与灾害监测预报的卫星，应用潜力巨大。但是其属于典型可见光-近红外影像，信息不足，且其CCD相机具有0~30°的扫描角，最大观测天顶角可达到35°，具有30m空间分辨率，这些设计特点导致HJ CCD影像无法直接应用传统的、较成熟的遥感影像处理算法。因此，开展山区HJ CCD影像气溶胶光学厚度与地表反射率反演具有现实需求和巨大的挑战。本研究针对山区的复杂地形因素和HJ CCD影像的设计特点，从辐射传输理论出发，开展山区HJ CCD影像气溶胶光学厚度和地表反射率反演的算法研究。 论文主要包括3方面研究内容：（1）山区HJ CCD影像气溶胶光学厚度反演。首先基于浓密森林像元在可见光波段反射率低的特点，引入红波段直方图阈值法自动提取山区可见光-近红外影像的浓密植被暗像元。然后，利用6S辐射传输模型建立查找表，基于红、蓝波段地表反射率间的线性关系，通过改进的暗目标法反演山区HJ CCD影像气溶胶光学厚度。（2）山区HJ CCD影像地表反射率反演。其中包括大气效应校正和地表邻近效应校正。基于臭氧和水汽辅助数据，提出结合改进暗目标法和辐射传输模型法的大气校正方法。其中以地表高程和观测天顶角为可变参数，利用辅助数据自定义大气模式，并通过反演的气溶胶光学厚度对查找表插值获得大气参数实现大气校正。在此基础上，基于三维辐射传输模型SHDOM引入大气点扩散函数的经验方程，结合大气辐射传输方程进行地表邻近效应校正，反演更准确的地表目标地物地表反射率。（3）HJ CCD影像气溶胶光学厚度和大气校正软件包实现。利用IDL可视化交互语言和ENVI遥感图像处理软件实现山区HJ CCD影像气溶胶光学厚度和地表反射率的相关算法，完成HJ CCD影像气溶胶光学厚度反演和大气校正自动化处理软件包。 论文主要得到以下几点结论：（1）红波段直方图阈值法能有效地提取山区可见光-近红外影像的浓密植被暗像元。该方法明显优于常用的可见光-近红外影像暗像元识别方法，识别精度接近传统的中红外波段影像的识别方法，提取结果能够满足暗目标法反演气溶胶光学厚度的要求。（2）改进暗目标法能有效地反演山区HJ CCD影像气溶胶光学厚度，HJ CCD影像反演结果与MODIS气溶胶产品的空间分布趋势和对比散点图的一致性非常好，且前者具有更高的空间分辨率特性，更适合山区气溶胶的遥感监测。（3）暗目标法与辐射传输模型法相结合的大气校正方法可以有效地反演山区HJ CCD影像地表反射率，实现大气校正。该方法能够消除大气效应影响，明显改善影像质量，使影像的光谱特征更接近真实值，其校正结果与TM影像地表反射率的一致性比较好。（4）利用三维辐射传输模型SHDOM引入的经验公式能有效的获取HJ CCD影像的背景反射率，进而实现地表邻近效应校正。邻近效应校正后影像的细节信息增多，纹理更加清晰，影像空间自相关性显著降低，邻近效应校正效果明显。（5）基于IDL和ENVI实现的HJ CCD影像气溶胶反演和大气校正软件包计算效率、自动化程度比较高，极大地提高了遥感影像的处理效率，具有较强的适用性。 本论文的研究为山区HJ CCD影像的自动化处理奠定了基础，为其定量化应用提供了可能。研究的创新点在于充分考虑HJ CCD影像的设计特征和山区环境对气溶胶光学厚度和地表反射率反演的影响，更有针对性地实现相关算法，并形成了自动化处理程序包。 |
| 英文摘要 | Both of the practical need of remote sensing application, and the development trend of quantitative remote sensing, decide remote sensing image can provide the surface reflectance information consistent with the actual characteristics of the surface. But the information received by remote sensing sensor will be affected mutually by atmospheric optical properties and surface environment radiation characteristics, which leads to the decrease of remote sensing image contrast, reduction of image quality, and the difficulty increase of surface information abstraction. Therefore, the atmospheric correction and adjacency effect correction of remote sensing image are indispensable and important links in the process of quantitative remote sensing for obtaining real surface reflectance. Aerosol has a strong scattering and absorption effect to the radiation of visible wave bands. It is the key atmospheric parameters in the atmospheric correction, and it has important role in the atmosphere radiation force and the atmospheric environmental quality assessment, etc. The aerosol retrieval from remote sensing has always been the hot spot of the quantitative remote sensing research. China is a mountainous country, and mountainous area accounts for about 70% of the nation's total. However mountainous terrain ups and downs, aerosol temporal and spatial change drastically, and climate changeful. and there are a lot of terrain shadow, cloud and shadow of cloud in mountain areas. All of these make the difficulty of areas aerosol optical thickness and surface reflectance retrieval increased. HJ CCD image is the satellite dedicated to environment monitoring and disaster prediction in our country, whose application potential is tremendous. But, HJ CCD image is typical visible and near infrared image whose information is insufficient, and it has 0~30°aspect angle, 0~35°observing zenith angle and 30m resolution. Because of these design features, HJ CCD image cannot be directly applied traditional remote sensing image processing algorithms, which is more mature.Therefore, there are realistic demand and great challenge for aerosol optical thickness and surface reflectance retrieval from HJ CCD image in mountain areas. Aiming at mountain complex terrain environment and the characteristics of HJ CCD image, the study of AOT and surface reflectance retrieval from HJ CCD image in mountain areas was carried out. The content of this parper includes 3 aspects: (1) Aerosol optical thickness retrieval from HJ CCD image in mountain area. Based on the low reflectance of forest pixels in visible bands, the red band histogram threshold method was proposed to automatically detect the DDV pixels from VIS-NIR images at mountainous areas. Subsequently, the lookup table was construted by 6s radiative transfer model and constant ratio between surface reflectance in red and blue bands was assumed, and then AOT of HJ CCD image was retrieved by an improved DDV method. (2) The retrieval of surface reflectance from HJ CCD image in mountain areas with the atmospheric correction and adjacency effect correction. Based on the auxiliary data of ozone and water vapour, the atmospheric correction method combined with improved DDV method and the radiative transfer model method was proposed. Then, the empirical equation of Atmospheric point spread function (PSF) was introduced from three dimensional radiative transfer model of SHDOM, and the adjacency effect correction was implemented, in order to obtain more accurate surface reflectance. (3) The completion of software development for AOT retrieval and atmospheric correction from HJ CCD image was achieved. All of the mentioned algorithm were realized by IDL language and ENVI software, and the processing automation of HJ CCD imaging AOT retrieval and atmospheric correction was Implement. The main conclution of this study are summarized as, (1) The red band histogram threshold method can automatically detected the DDV pixels from VIS-NIR im |
| 语种 | 中文 |
| 内容类型 | 学位论文 |
| 源URL | [http://ir.imde.ac.cn/handle/131551/8008]  |
| 专题 | 成都山地灾害与环境研究所_数字山地与遥感应用中心 |
| 推荐引用方式 GB/T 7714 | 赵志强. 山区HJ CCD影像气溶胶光学厚度与地表反射率反演算法研究[D]. 北京. 中国科学院研究生院. 2014. |
| 个性服务 |
| 查看访问统计 |
| 相关权益政策 |
| 暂无数据 |
| 收藏/分享 |
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。
修改评论