基于证据理论的遥感图像变化检测理论与方法研究

项目名称： 基于证据理论的遥感图像变化检测理论与方法研究

项目编号： No.61203222

项目类型： 青年科学基金项目

立项/批准年度： 2013

项目学科： 自动化学科

项目作者： 杨艺

作者单位： 西安交通大学

项目金额： 24万元

中文摘要： 遥感图像变化检测技术是基于同一地域不同时相的遥感图像信息，获取所关注对象的变化信息并评估其时空分布规律的技术。遥感图像变化检测具有重要的理论研究意义和重大的应用价值。遥感图像变化检测从数据获取开始，一直伴随着不确定性，这也制约着遥感图像变化检测的进一步发展。因此提升遥感图像变化检测质量，首先要对不确定信息进行有效建模与处理。作为一种不确定信息融合技术，证据理论具有强大的不确定信息表示与融合能力，但直接应用于遥感领域中存在一定的困难。本项目基于证据理论对遥感图像变化检测各个环节中的多源不确定信息分别进行建模与处理，包括图像配准、特征提取、地物分类、变化判定及性质鉴别等，并结合各环节的实际对证据理论的不足进行改进。针对上述研究问题取得的进展和突破，能从理论和应用层面完善与丰富遥感图像变化检测理论与方法，对提升遥感信息解译能力以及拓展证据理论这一不确定信息融合技术的应用范围具有重大作用和意义。

中文关键词： 证据理论；遥感图像；变化检测；图像配准；

英文摘要： Change detection of remote sensing images is a process of acquiring changes and estimating temporal-spatial distribution discipline from objects of interest based on remotely sensed imagery of the same scene observed at different times. It has important theoretical significance and application value. Uncertainty accompanies change detection of remote sensing images from the beginning of data acquisition step, which constrains the further development of change detection. Consequently, in order to improve the change detection quality, effective uncertainty information modeling and processing are firstly needed. As an uncertainty information fusion technique, evidence theory has powerful uncertainty information representation and fusion ability, but there still exist certain difficulties in its direct application in remote sensing area. In this project multi-source uncertainty information in each part of remote sensing images change detection, including image registration, feature extraction, terrain classification, change determination and change type identification, etc. is respectively modeled and processed based on evidence theory. Weakness of evidence theory is also improved concerning actual situations in each part. Progress and breakthroughs obtained for the aforementioned research problems can perfect and e

英文关键词： Evidence Theory；Remote Sensing Images；Change Detection；Image Registration；