国产高分卫星影像质量评价及特征分析

发布时间：2018-08-01 17:54

【摘要】：高分辨率对地观测系统重大专项是《国家中长期科学与技术发展规划纲要(2006-2020年)》确定的十六个重大科技专项之一,于2010年批准启动实施,目前已经建立了较为完善的地面系统。高分一号(GF-1)是该专项首颗卫星,于2013年成功发射,高分二号(GF-2)是我国自主研制的首颗空间分辨率优于1m的民用光学遥感卫星,标志着我国遥感卫星进入亚米级“高分时代”。卫星遥感影像因其自身具有的客观、宏观、快速的对地观测能力,已成为国土资源规划、决策与管理的重要支撑性技术之一。数据的质量决定了信息的可用性,对所成影像进行质量评价是卫星业务化运行和应用的重要保证。通过对卫星影像质量评价和特征分析,得到图像处理和信息提取的先验知识,有助于高分卫星影像的推广利用和后继卫星的开发研制,这也正是本研究的出发点和落脚点。本文依托四川省国土资源厅科学研究计划项目“国产卫星数据在土地利用监测与现状变更调查中的遥感产品验证与推广研究”,重点针对GF-2影像数据开展质量评价与影像特征分析,研究结果可为高分卫星的应用与发展提供参考,并为充分挖掘高分系列卫星应用价值奠定基础,主要工作及取得的成果如下:(1)对影像处理中常用的波段组合、融合方法以及几何校正进行试验分析,得到了最优波段组合、最佳融合方法和影像几何校正精度,分析了GF-2影像的成图比例尺,试验结果可为GF-2数据应用提供参考;(2)选取覆盖相同区域且成像时间较为接近的GF-1、GF-2卫星影像,以影像公共区域为研究区,从地理要素视觉分析、灰度统计、信息量、清晰度、影像纹理特征以及影像能量特征等方面开展影像工程质量分析和评价,定性和定量评价了GF-2卫星影像的成像质量;(3)从信息提取的角度,对比分析GF-1、GF-2卫星的多光谱影像典型地物光谱特征,结合地物样本可分性测试和监督分类试验,验证了GF-2卫星获取地物信息的能力;(4)调制传递函数(MTF)是非常精确的光学系统性能评价指标,也常常用来评价图像质量。本文利用刀刃法获取了GF-2全色影像的调制传递函数(MTF),通过分析GF-2全色影像的FWHM值和Nyquist频率处MTF值,定量评价了GF-2卫星全色相机的空间分辨率。
[Abstract]:The high resolution Earth observation system is one of the 16 major scientific and technological projects defined in the outline of the National Medium-and Long-Term Science and Technology Development Program (2006-2020). It was approved for implementation in 2010, and a relatively perfect ground system has been established. Gao Fei-1 (GF-1) is the first satellite of this project, which was successfully launched in 2013. Gaofen 2 (GF-2) is the first civil optical remote sensing satellite with spatial resolution better than 1m, which marks the entry of China's remote sensing satellite into the sub-meter "high score era". Satellite remote sensing image has become one of the important supporting technologies in land and resource planning, decision-making and management because of its objective, macroscopic and rapid earth observation ability. The quality of data determines the availability of information. The evaluation of the image quality is an important guarantee for satellite operation and application. A priori knowledge of image processing and information extraction is obtained through the quality evaluation and feature analysis of satellite images, which is helpful to the popularization and utilization of high-score satellite images and the development and development of subsequent satellites, which is the starting point and destination of this study. This paper relies on the scientific research project of Sichuan Provincial Land and Resources Department, "validation and popularization of remote sensing products of domestic satellite data in land use monitoring and status change investigation", Focusing on the quality evaluation and image feature analysis of GF-2 image data, the research results can provide a reference for the application and development of high-score satellites, and lay a foundation for fully mining the application value of high-score series satellites. The main work and achievements are as follows: (1) the band combination, fusion method and geometric correction used in image processing are tested and analyzed, and the optimal band combination, optimal fusion method and image geometric correction accuracy are obtained. The mapping scale of GF-2 image is analyzed. The experimental results can provide reference for the application of GF-2 data. (2) selecting GF-1 and GF-2 satellite images covering the same area and the same imaging time, taking the common area of the image as the study area, the visual analysis of geographical elements is carried out. Image engineering quality analysis and evaluation are carried out in the aspects of grayscale statistics, information quantity, clarity, texture features and image energy features. The image quality of GF-2 satellite images is evaluated qualitatively and quantitatively. (3) from the angle of information extraction, The spectral characteristics of typical ground objects in multispectral images of GF-1 and GF-2 satellite are compared and analyzed, combined with the test of the separability of ground objects and the supervised classification test. (4) Modulation transfer function (MTF) is a very accurate index for evaluating the performance of optical systems and is often used to evaluate image quality. In this paper, the modulation transfer function (MTF),) of GF-2 panchromatic images is obtained by using the blade method. The spatial resolution of GF-2 satellite panchromatic cameras is quantitatively evaluated by analyzing the FWHM values of GF-2 panchromatic images and the MTF values at Nyquist frequencies.
【学位授予单位】：成都理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P237

【参考文献】