| 多先验形状模型变分水平集建筑物检测 |
| |
| 引用本文： | 杨剑,田昊.多先验形状模型变分水平集建筑物检测[J].科学技术与工程,2010,10(4). |
| |
| 作者姓名： | 杨剑 田昊 |
| |
| 作者单位： | 国防科学技术大学信息系统与管理学院,长沙,410073 |
| |
| 摘 要： | 提出了针对城市遥感影像中建筑物目标分割的一种新的方法。该方法在简化Mumford-Shah模型的水平集方法(Chan-Vese模型方法)的基础上,加入建筑物先验形状模型信息,并引入标记函数来动态选择先验形状,最后通过推演新的模型更好地实现了多个建筑物目标的分割。
|
| 关 键 词： | 图像分割 城市遥感影像 Chan-Vese模型 先验形状模型 标记函数 |
| 收稿时间： | 2009/10/27 0:00:00 |
| 修稿时间： | 2009/11/18 0:00:00 |
The building detection from remote sensing image based on variational level set method and multiple prior shape |
| |
| YANG Jian and TIAN Hao.The building detection from remote sensing image based on variational level set method and multiple prior shape[J].Science Technology and Engineering,2010,10(4). |
| |
| Authors: | YANG Jian and TIAN Hao |
| |
| Institution: | Institute of Information System and Management/a>;National University of Defense Technology/a>;Changsha 410073/a>;P.R.China |
| |
| Abstract: | A novel method for detecting multiple building objects from remote sensing image is described.The method is based on a typical simplified Mumford - Shah model(Chan-Vese model),and multiple prior shape models and labeling function which can dynamicaUy choose prior shape are imported,finally the new formed model is e-volved and it is used in the multiple building objects detection and a better result is obtained. |
| |
| Keywords: | image segmentation city remote sensing Chan-Vese model prior shape model labeling function |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
| 点击此处可从《科学技术与工程》浏览原始摘要信息 |
| 点击此处可从《科学技术与工程》下载免费的PDF全文 |
