CASIA OpenIR  > 毕业生  > 硕士学位论文
Alternative TitleResearch on Change Detection Using Remote Sensing Images
Thesis Advisor卢汉清
Degree Grantor中国科学院研究生院
Place of Conferral中国科学院自动化研究所
Degree Discipline模式识别与智能系统
Keyword变化检测 合成孔径雷达 机器学习 纹理特征 Chang Detection Sar Machine Learning Texture Feature
Abstract由于变化检测技术具有非常广泛的应用背景,在诸如城区规划、植被覆盖调查、防灾减灾、地图更新等方面都有着相当广泛的应用,因此一直是遥感图像处理中的研究热点之一。本文在总结了目前已有的遥感图像变化检测方法的基础上,提出了两种针对特定场景的遥感图像变化检测方法,并实现了一个城市场景下的变化检测系统。本文的主要工作归纳如下: (1)针对各种变化检测方法进行了综述,分析了这些方法的优缺点以及适用范围,对于后续工作中在遥感图像中提取合适的特征以及提出新的方法提供了参考与借鉴。 (2)在合成孔径雷达图像中提取区域内多种特征,并将机器学习中的Adaboost算法引入到特征选择中,通过对于训练集图像中变化的象素点和没有变化的象素点的学习,提出了一种对于合成孔径雷达图像中斑点噪声较为鲁棒的变化检测方法。 (3)通过对于城市场景的遥感图像的分析,提出了一种利用特征线段以及颜色纹理信息的多特征结合的城市场景下变化检测方法。该方法通过边缘信息、区域纹理信息为依据来判断是否发生了真实的变化。使用该方法实现的城市场景下的变化检测系统能够对于城市中外观、颜色发生明显变化的较大区域进行自动的检测。
Other AbstractChange Detection is one of the most popular research fields in Remote Sensing Image Processing. It has broad applications in land use planning, vegetation overlay investigate, damage prevention and map updating. In this dissertation, two specific methods are proposed, and we build up a change detection system for urban changes. The main contributions are as follows: (1)Based on a brief review of the existing change detection methods, their advantages and disadvantages as well as application situation are compared, which will help us to extract appropriate features and put up new methods. (2)We use the texture features extracted from SAR images. To select the appropriate features, the Adaboost algorithm is adopted. Based on the changed and unchanged pixels in the training set images, we present a change detection method that is quite robust to the speckle noise that is common in SAR images. (3)Based on remote sensing images of the city zone, we present a change detection method, which uses edge line and color information as the features in urban area. This approach uses the edge and color information to identify changes between two remote sensing images. Based on this method, we build up a real change detection system to detect changed regions in remote sensing images, which can find relatively obvious changed regions in city automaticlly.
Other Identifier200328014604158
Document Type学位论文
Recommended Citation
GB/T 7714
张世清. 遥感图像变化检测研究[D]. 中国科学院自动化研究所. 中国科学院研究生院,2006.
Files in This Item:
File Name/Size DocType Version Access License
CASIA_20032801460415(3220KB) 暂不开放CC BY-NC-SAApplication Full Text
Related Services
Recommend this item
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[张世清]'s Articles
Baidu academic
Similar articles in Baidu academic
[张世清]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[张世清]'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.