| 基于面向对象的滑坡快速识别技术研究 |
| |
| 引用本文： | 周志华,林维芳,许高程,谢先明,王松.基于面向对象的滑坡快速识别技术研究[J].安徽农业科学,2012(5):3017-3018,3071. |
| |
| 作者姓名： | 周志华 林维芳 许高程 谢先明 王松 |
| |
| 作者单位： | 1. 广东省工程勘察院,广东广州,510510
2. 北京全景天地科技有限公司,北京,100085 |
| |
| 摘 要： | 对滑坡快速识别技术研究,可为滑坡灾害调查、灾后快速反应以及道路选线等提供重要依据。利用遥感影像,采用面向对象的影像分析技术与多尺度影像分割技术,运用eCognition软件对日本广岛地区进行滑坡快速识别。针对研究区复杂的地物类型,构建优化分类模型,选择适宜的影像分割尺度和分类特征,按照合理顺序逐步识别滑坡体。试验表明,该方法识别速度快、精度高,为遥感影像滑坡快速识别提供了新思路。
|
| 关 键 词： | 面向对象 多尺度分割 滑坡识别 |
Research of Fast Landslide Recognition Based on Object-oriented Technology |
| |
| Institution: | ZHOU Zhi-hua et al(Guangdong Engineering Investigation Institute,Guangzhou,Guangdong 510510) |
| |
| Abstract: | The research of fast landslide recognition provides important bases for landslide investigation,fast reaction after disaster and road route selection.Using remote sensing images and object-oriented and multi-resolution image segmentation and adopting the eCognition software,the study conducted a research into Hiroshima region.According to the different types of land in the city,the researchers created the module of classification,selected proper scale parameter and classification feature,and then took proper sequence to fast recognise landslide.In addition,the study also provided a universal rule for landslide recognition.The test indicated that this method has a precise result and short course,providing a new way for the research on fast landslide recognition from remote sensing image. |
| |
| Keywords: | Object-oriented Multi-resolution segmentation Landslide recognition |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
