敏捷卫星遥感图像配准和拼接技术研究
发布时间：2019-01-02 07:25
【摘要】：遥感图像是指从远距离平台上利用光电成像载荷获取的地物目标图像,常见的平台如飞机和卫星。由于其覆盖范围广,光谱频段丰富,在军事和民用上都起到了巨大的作用。但是随着人类活动的区域日益广泛,对于遥感图像的整体要求也越来越高,希望在获得高分辨率图像的同时,其覆盖的区域更加的宽广,以至于能对一个较大的区域进行细致深入的分析研究,为后续的人类决策提供支持。但是,受限于目前的技术水平,高分辨率和宽视场仍然是一对矛盾体。常见的解决方法是通过图像配准和拼接的方法,也就是是利用传感器得到多幅高分辨率的小视场图像,然后对这些具有一定重叠区域的图像进行配准拼接,得到一幅大视场图像。这种矛盾在敏捷卫星的任务中会出现的更加频繁,敏捷卫星是近年来出现的新型卫星,具有优秀的姿态加速功能,使得敏捷卫星可以对目标进行更快速的瞄准和更精确的扫描,大大提高了满足复杂任务的要求。但是敏捷卫星由于其高机动性,会使得成像环境更加的复杂,从而影响最终的遥感图像质量,因此对敏捷卫星的成像模式进行深入细致地研究,可以定量化地分析机动成像后的图像质量变化特性,而因为敏捷特性带来的图像配准拼接问题也会成为更大的挑战。 通过对经典的配准和拼接算法的研究,为后续的敏捷卫星的图像拼接约束分析提供理论支持。图像配准拼接技术可分为两类：基于区域的图像配准技术,如序列相似性检测算法、交叉相关相似性度量函数和傅里叶变换算法；基于特征点配准算法包括sIFT特征点提取算法和Harris角点提取算法。对这些配准算法的研究可以从图像角度对敏捷模式带来的影响进行分析,同时也为基于遥感景物内容特性的配准算法提供了基础。 通过构建几何模型、辐射模型、相机模型和大气模型,来模拟分析敏捷卫星的机动成像模式,并且提出梯度信息熵、仿射退化度和立体结构相似度等新颖评价指标来预估图像质量。基于四个模型对敏捷卫星的成像模式进行分解,使得可以通过卫星轨道等基本参数来定量地分析卫星图像的各个参数,比如分辨率、幅宽,以及后续的图像质量指标。在理论模型的基础上,编写了“敏捷卫星成像仿真和质量分析“软件,可以对敏捷卫星进行各个模型的分析,同时可以对图像退化进行仿真模拟。 利用图像复原和图像融合的预处理技术,改善敏捷成像下可能带来的图像模糊和颜色退化现象。针对图像复原技术,首先分析了典型的复原算法,然后提出了基于FOE模型的图像复原算法,最后研究对比了图像复原技术对于特征点提取技术的影响。针对图像颜色退化,本文分析了典型的图像融合算法,并对融合后的遥感图像进行配准拼接,提高其目标识别的能力。 通过对遥感图像景物内容特性的分析,提出了几种改善性的配准和拼接算法,使得遥感图像有更广泛的实际应用能力。首先提出了一种基于梯度信息权重优化的配准技术,利用图像的梯度信息对图像的特征点进行权重划分,利用权重值对特征点进行区别优化,最后将联合的图像拼接技术应用到重叠区域,实现对敏感区域进行高精度匹配的目的。针对遥感图像含有丰富的内容特性,设计了一种双特征点配准算法,利用SIFT和Harris短发提取特征点种类的区别,分别对角点区域密集的地方实现Harris特征点提取,对于相对平坦的区域(比如草地和水面)进行SIFT特征点提取,最后得到可以应付复杂环境下得到的推扫图像。遥感图像通常含有丰富的内容,使得其特征点数量通常非常的巨大,应对这种情况,提出了基于精炼控制点的配准技术,提取少量的精确特征点可以减少错误匹配特征点对结果的干扰,同时可以降低对遥感图像的处理难度。
[Abstract]:Remote sensing images refer to figure object images, common platforms such as aircraft and satellites, acquired from a remote platform using a photo-imaging load. Because of its wide range of coverage, the spectral band is rich and plays a great role in both military and civilian use. But as the area of human activity is becoming more and more extensive, the overall requirement of the remote sensing image is higher and higher, and it is hoped that the area covered by the remote sensing image is wide enough to carry out detailed and in-depth analysis and research on a large area, Support for follow-up human decision-making. However, limited to the current state of the art, the high resolution and wide field of view are still a pair of contradictions. The common solution method is to obtain a large-field-of-view image by using a sensor to obtain a plurality of high-resolution small-field-of-view images by using a sensor to obtain a plurality of high-resolution small-field-of-view images. The contradiction is more frequent in the task of the agile satellite, and the agile satellite is a new type of satellite that has emerged in recent years, has excellent attitude acceleration function, so that the agile satellite can carry out more rapid aiming and more accurate scanning of the target, and the requirements of meeting complex tasks are greatly improved. However, because of its high mobility, the agile satellite can make the imaging environment more complex, so as to influence the quality of the final remote sensing image, so that the imaging mode of the agile satellite is deeply researched, and the image quality change characteristic after the mobile imaging can be quantitatively analyzed, The image matching problem caused by the agile character can also become a bigger challenge. Through the research of the classical alignment and stitching algorithm, the paper provides the theoretical support for the analysis of the image mosaic restriction of the subsequent agile satellite. The image matching technique can be divided into two types: region-based image matching technique, such as sequence similarity detection algorithm, cross-correlation similarity measure function and Fourier transform algorithm, and feature point matching algorithm including the SIFT feature point extraction algorithm and the Harris corner point extraction algorithm. In this paper, the influence of the image angle on the agile model can be analyzed, and the basis of the alignment algorithm based on the content characteristics of the remote sensing scene is also provided. By building the geometric model, the radiation model, the camera model and the atmosphere model, the maneuvering imaging mode of the agile satellite is simulated, and the new evaluation indexes such as the gradient information entropy, the affine degradation degree and the three-dimensional structure similarity are put forward to estimate the map. The imaging mode of the agile satellite is decomposed based on the four models, so that the parameters of the satellite image can be quantitatively analyzed by the basic parameters such as satellite orbit, such as resolution, width, and subsequent image quality. On the base of the theoretical model, the 鈥淪imulation and quality analysis of agile satellite imaging鈥，
本文编号：2398187
[Abstract]:Remote sensing images refer to figure object images, common platforms such as aircraft and satellites, acquired from a remote platform using a photo-imaging load. Because of its wide range of coverage, the spectral band is rich and plays a great role in both military and civilian use. But as the area of human activity is becoming more and more extensive, the overall requirement of the remote sensing image is higher and higher, and it is hoped that the area covered by the remote sensing image is wide enough to carry out detailed and in-depth analysis and research on a large area, Support for follow-up human decision-making. However, limited to the current state of the art, the high resolution and wide field of view are still a pair of contradictions. The common solution method is to obtain a large-field-of-view image by using a sensor to obtain a plurality of high-resolution small-field-of-view images by using a sensor to obtain a plurality of high-resolution small-field-of-view images. The contradiction is more frequent in the task of the agile satellite, and the agile satellite is a new type of satellite that has emerged in recent years, has excellent attitude acceleration function, so that the agile satellite can carry out more rapid aiming and more accurate scanning of the target, and the requirements of meeting complex tasks are greatly improved. However, because of its high mobility, the agile satellite can make the imaging environment more complex, so as to influence the quality of the final remote sensing image, so that the imaging mode of the agile satellite is deeply researched, and the image quality change characteristic after the mobile imaging can be quantitatively analyzed, The image matching problem caused by the agile character can also become a bigger challenge. Through the research of the classical alignment and stitching algorithm, the paper provides the theoretical support for the analysis of the image mosaic restriction of the subsequent agile satellite. The image matching technique can be divided into two types: region-based image matching technique, such as sequence similarity detection algorithm, cross-correlation similarity measure function and Fourier transform algorithm, and feature point matching algorithm including the SIFT feature point extraction algorithm and the Harris corner point extraction algorithm. In this paper, the influence of the image angle on the agile model can be analyzed, and the basis of the alignment algorithm based on the content characteristics of the remote sensing scene is also provided. By building the geometric model, the radiation model, the camera model and the atmosphere model, the maneuvering imaging mode of the agile satellite is simulated, and the new evaluation indexes such as the gradient information entropy, the affine degradation degree and the three-dimensional structure similarity are put forward to estimate the map. The imaging mode of the agile satellite is decomposed based on the four models, so that the parameters of the satellite image can be quantitatively analyzed by the basic parameters such as satellite orbit, such as resolution, width, and subsequent image quality. On the base of the theoretical model, the 鈥淪imulation and quality analysis of agile satellite imaging鈥，
本文编号：2398187
本文链接：https://www.wllwen.com/guanlilunwen/gongchengguanli/2398187.html
