EGI OpenIR  > 研究系统  > 荒漠环境研究室
Thesis Advisor周可法
Degree Grantor中国科学院大学
Place of Conferral北京
Degree Discipline测绘工程
Keyword成像高光谱遥感 图像去噪 大气校正 Smile校正 几何精校正
Abstract高光谱探测作为一种有效的现代化探测技术,能够提供被探测对象全面、真实的空间信息与光谱信息,近年来,被大范围的应用于军事及民用领域。论文以HySpex机载成像光谱仪数据作为主要数据资料,深入研究了高光谱图像处理过程中几个较为关键的环节,并制定了有效的解决措施, 为高光谱图像的应用研究及数据处理工作提供了理论前提和基础。 论文的主要研究内容包括: (1)对高光谱图像的噪声特点及其来源进行了全面、系统的分析,阐述了当前业内常用的、较为有效的滤波手段,并对比分析了光谱域随机噪声与条带噪声;论文以HySpex影像中条带噪声所具有的高频特征为依据,采用一种行平滑滤波计算方法来有效地去除噪声所带来的影响;对比分析各类噪声去除方法,设计了曲率差分自适应全变分去噪算法;利用空间-光谱维的相关信息,研究了NSCT和KPCA相结合的高光谱图像去噪方法,应用在HySpex数据处理中有较好的效果。 (2)采用QUAC与FLAASH模型对HySpex数据进行大气校正,对比分析校正前后的影像信息,提出了BCLQUAC方法,进行薄云影像的快速大气校正,能够有效保证大气校正的纯度与精度,且速度远快于QUAC模型。针对HySpex成像光谱仪中典型的光谱失真歪曲问题, 对Smile效应进行了分析,采样MNF空间列均值调节方法,解决了HySpex数据中光谱特征的扭曲问题。 (3)论文分析航空定位导航姿态测量系统(IMU/DGPS)的数据,通过大量的计算试验,厘定系统计算参数,完成高光谱图像的几何精校正;利用传感器的多个参数对每个像元的地面坐标进行计算,重新采样原始图像像元,最终得到校正图像。通过实地验证和精度评价,该几何精纠正方案能够取得较好的纠正效果。
Other AbstractHyperspectral detection is an effective type of remote sensing technology with high spectral resolution, which provides abundant spatial and spectral information about the observed object, and plays an important role in many fields of military, civil and other application. HySpex airborne imaging spectrometer data were used in this paper, and some key procedures in the hyperspectral image processing were further researched in detail, and effective solutions were proposed. The results of study provide theoretical basis for the data processing and applications of hyperspectral images. The followings are the summary of our contributions. (1) The noise’s characteristics and sources of the hyperspectral image are systematically analyzed, several common filter methods are discussed. Comparisons are made by filter experiments with spectral domain random noise and band noise. Basing on the high frequency characteristics of stripe noise of HySpex images, we used smoothing filter calculation mode to eliminate the influences caused by noise. Based on the analysis of previous methods of removing random noises of images,this paper develops an adaptive total variation denoising algorithm based on curvature differential removing random noises from imaging spectrometer images.Then a hyperspectral image denoising method based on nonsubsampled contourlet transform (NSCT) and keneral principal component analysis (KPCA) is given. (2) The QUAC and FLAASH are used to as the atmospheric correction models, and the HySpex data is used as the original data to analyze the precision of atmospheric corretation. The BCLQUAC atmospheric correction model is proposed, and it is faster and better to HySpex data than the QUAC model. With the consideration of spectral distortions are generally existed in the HySpex imaging spectrometer, smile effect was introducedand MNF method was used to solve spectral distortions of HySpex images. (3) This paper investigated the process of geometric correction of hyperspectral images directly using IMU/DGPS data. We used parameters of roll angle sensors to calculateground coordinate and and resampled the original image pixel to obtain the corrected images. The result showed that our proposed geometric correction scheme can achieve good results.
Subject Area测绘工程
Document Type学位论文
Recommended Citation
GB/T 7714
刘盈娣. HySpex成像高光谱图像处理关键技术研究[D]. 北京. 中国科学院大学,2015.
Files in This Item:
There are no files associated with this item.
Related Services
Recommend this item
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[刘盈娣]'s Articles
Baidu academic
Similar articles in Baidu academic
[刘盈娣]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[刘盈娣]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.