两种快速星像匹配算法的比较

星像的位置和亮度等信息对天文研究和空间导航具有基础意义。人们常常需要快速匹配从密集星场CCD图像中提取的信息来获取与星表中对应的信息(如位置、光度等),因此开发出了多种星像匹配算法。重点比较了两种最新的星像匹配算法:i)基于向量的方法;ii)基于径向和环向特征的方法。通过对云南天文台1m望远镜观测的CCD图像进行实际匹配,结果表明基于向量的方法更优,它是一种快速的、接近常数时间的匹配算法。     

基于Retinex算法的数字航空影像色彩校正研究

针对轻小型无人机数字航空影像的色彩校正问题,笔者采用先匀光,后校正,将原始影像分块,每一块影像都对R、G、B三个通道进行处理,分别在不匀光、Mask匀光、Wallis匀光和Retinex匀光下,基于近距离拍摄的参考影像进行色彩校正,最后合并3个通道及所有分块,消除拼接缝.通过定性分析与定量分析的方法,对比不匀光及3种匀...     

顾及粗差的低空数字影像拼接

低空遥感影像相对于传统航空和卫星影像覆盖面积小且像幅数多,影像拼接是内业处理的重要工作之一。低空摄影机畸变大及飞行平台不稳定等因素易产生同名点匹配定位粗差,可能致使拼接低空影像接边误差较大。为减少这些粗差,利用具有尺度、旋转和平移不变性的SIFT算法进行低空影像匹配,然后采用Huber算法约束粗差,并通过Levenberg-Marquardt非线性最小二乘法进行平差,以获得精确的影像拼接单应矩阵。实验验证了该方法可减少拼接缝,提高了拼接精度。     

基于GPS/IMU的红外影像直接定位技术

研究无控制点情况下的红外影像目标快速定位技术,有利于发挥红外遥感的测绘应用潜力。首先介绍了基于GPS/IMU的目标定位原理与方法;然后实验验证了红外影像多像前方交会和GPS/IMU辅助红外影像光束法区域网平差的可行性;并分析了误差产生的原因。实验证明,航空红外遥感在GPS/IMU组合系统的配合下,能满足目标探测与预警的精度要求,具有实用价值。     

SPOT 5卫星影像测绘能力分析

HRS同轨立体影像获取能力是SPOT5卫星最突出的一个关于测绘制图的新特性,它确定了SPOT5卫星在测绘中的重要应用前景。本文从SPOT5HRS的立体成像方式,SPOTImage公司最新推出的Reference3D产品及陕西宝鸡地区的HRS立体影像测量试验等方面对SPOT5HRS影像测绘能力进行了分析和评价,说明了SPOT5卫星影像测绘将是解决我国西部困难地区基础测绘的一种重要手段。     

2007年7月1 7日中天山暴雨天气过程分析

通过对2007年7月17日中天山暴雨天气过程的分析,总结了这次暴雨天气发生的有利环流形势、物理量场分布、卫星云图特征及雷达回波特征,为此类暴雨天气预报提供了参考依据。     

OMIS图像几何校正研究

与卫星遥感传感器相比,机载成像光谱仪具有平台姿态稳定性差、飞行高度低、视场角较大等特点,这些因素的结合使图像几何畸变的图案复杂化,几何纠正的目的就是尽可能消除这些几何畸变。该文详细分析了OMIS图像存在的主要几何畸变,提出OMIS图像几何校正的处理流程,试验结果表明该方法能取得较好的校正效果。     

Modelling the temporal variation in snow-covered area derived from satellite images for simulating/forecasting of snowmelt runoff in Turkey/Modélisation de la variation temporelle de la surface enneigée à partir d'images satellitaires pour la simulation/prévision de l'écoulement de fonte nivale en Turquie

Abstract

Monitoring the change of snow-covered area (SCA) in a basin is vitally important for optimum operation of water resources, where the main contribution comes from snowmelt. A methodology for obtaining the depletion pattern of SCA, which is based on satellite image observations where mean daily air temperature is used, is applied for the 1997 water year and tested for the 1998 water year. The study is performed at the Upper Euphrates River basin in Turkey (10 216 km²). The major melting period in this basin starts in early April. The cumulated mean daily air temperature (CMAT) is correlated to the depletion of snow-covered area with the start of melting. The analysis revealed that SCA values obtained from NOAA-AVHRR satellite images are exponentially correlated to CMAT for the whole basin in a lumped manner, where R ² values of 0.98 and 0.99 were obtained for the water years 1997 and 1998, respectively. The applied methodology enables the interpolation between the SCA observations and extrapolation. Such a procedure reduces the number of satellite images required for analysis and provides solution for the cloud-obscured images. Based on the image availability, the effect of the number of images on the quality of snowmelt runoff simulations is also discussed. In deriving the depletion curve for SCA, if the number of images is reduced, the timing of image analysis within the snowmelt period is found very important. Analysis of the timing of satellite images indicated that images from the early and middle parts of the melt period are more important.     

An approach to the radiometric aerotriangulation of photogrammetric images

Harnessing the radiometric information provided by photogrammetric flights could be useful in increasing the thematic applications of aerial images. The aim of this paper is to improve relative and absolute homogenization in aerial images by applying atmospheric correction and treatment of bidirectional effects. We propose combining remote sensing methodologies based on radiative transfer models and photogrammetry models, taking into account the three-dimensional geometry of the images (external orientation and Digital Elevation Model). The photogrammetric flight was done with a Z/I Digital Mapping Camera (DMC) with a Ground Sample Distance (GSD) of 45 cm. Spectral field data were acquired by defining radiometric control points in order to apply atmospheric correction models, obtaining calibration parameters from the camera and surface reflectance images. Kernel-driven models were applied to correct the anisotropy caused by the bidirectional reflectance distribution function (BRDF) of surfaces viewed under large observation angles with constant illumination, using the overlapping area between images and the establishment of radiometric tie points. Two case studies were used: 8-bit images with applied Lookup Tables (LUTs) resulting from the conventional photogrammetric workflow for BRDF studies and original 12-bit images (Low Resolution Color, LRC) for the correction of atmospheric and bidirectional effects. The proposed methodology shows promising results in the different phases of the process. The geometric kernel that shows the best performance is the Lidense kernel. The homogenization factor in 8-bit images ranged from 6% to 25% relative to the range of digital numbers (0–255), and from 18% to 35% relative to levels of reflectance (0–100) in the 12-bit images, representing a relative improvement of approximately 1–30%, depending on the band analyzed.