多源遥感图像融合技术及地物信息提取研究.docx

 

1.94万字             53页

摘 要

多源遥感图像融合是图像融合中的一个重要分支部分。现代信息时代的今天，多源遥感图像融合已成为图像内容和图像处理领域中必不可少的技术，它在很多军用的方面和民用方面有着举足轻重的地位。

    随着传感器技术的发展，遥感技术为对地观测提供的遥感图像数据越来越丰富。遥感图像融合技术可以将多源遥感数据所包含的信息优势或互补性有机地结合起来，并且能有效的提高遥感图像的空间分辨率，增强图像的特征性，提高分类精度和动态监测的能力。本文介绍了图像处理的几种方法，研究了遥感图像的融合方法。对图像的数据进行融合处理可以提高图像的分辨率、图像的准确度和置信度。本文的主要研究内容和工作总结如下：

1、介绍了图像融合中的基本原理，研究了图像预处理过程图像增强、图像配准等问题，为开展图像融合做准备。

2、研究了传统的图像融合方法：IHS变换和PCA变换，将两种融合方法应用到Landsat 8卫星多光谱与全色图像融合之中。研究结果表明：经过两种融合方法后融合图像更加清晰，分辨率明显提高，图像的信息量更加丰富。从相关系数、偏差定量指标来看PCA变换优于HIS变换。从平均梯度、交叉熵和熵定量指标来看IHS变换优于PCA变换。

3、针对IHS变换和PCA变换融合方法存在的光谱扭曲，图像的清晰度不够等问题，研究了基于静态小波变换（SWT）的图像融合方法。结果表明：从相关系数、信息熵、偏差和交叉熵定量指标来看，SWT变换明显优于IHS、PCA变换。

关键词：图像融合、遥感、IHS变换、PCA变换、质量评价

ABSTRACT

Multi-source remote sensing image fusion is an important part of image fusion technology, and it is essential for image content preservation and image analysis. It has frequently been utilized in military and civil fields.

With the development of remote sensing sensor technology, remote sensing has provided more and more abundant remote sensing image data so as to observe earth surface. Image fusion technique can combine multi-source information, and can effectively improve the spatial resolution and enhance image features. Therefore, it can improve classification accuracy and promote dynamic monitoring. This paper describes image pre-processing and several image fusion methods. The summary is as follows:

1. The paper introduced the basic theory about image enhancement, image registration and image fusion.

2. The paper studied traditional image fusion method, such as IHS transform and PCA transform, and then utilized them to Landsat 8 multi-spectral and panchromatic image fusion processing. The experimental results had shown that the pan-sharpened images were clearer than original multi-spectral images, moreover, they had higher spatial resolution and more abundant image information. From correlation coefficient and deviation quantitative indexes point of view, it was shown that PCA transform was better than HIS transform, and from mean gradient, cross entropy and entropy indexes point of view, the IHS transform was better than PCA transform.

3. For solving spectral distortion introduced by IHS transform and PCA transform, and not enough image definition, the static wavelet transform (SWT) was proposed in this paper. The experimental results had shown that from correlation coefficient, information entropy, deviation and cross entropy indexes point of view, SWT transform was obviously better than IHS and PCA transform.

Keywords: image fusion, remote sensing, IHS transform, PCA transform, quality assessment.