面向对象的高空间分辨率遥感影像信息提取——汶川地震城市震害房屋案例研究
|
摘要
随着遥感技术的快速发展,尤其是高分辨率民用遥感卫星的成功发射和应用,遥感技术在建筑物识别、震灾调查和快速预估评估方面发挥出越来越重要的作用。然而,传统面向像元的分类方法在对高分辨率遥感影像进行分类时,存在着不能充分利用影像信息、分类精度降低、速度慢等局限性,根据高分辨率遥感影像的特点,提出了面向对象的高空间分辨率遥感影像建筑物震害信息提取方法。该方法首先通过影像分割将影像划分为互不相交的影像对象,然后根据这些影像对象的影像特征如光谱、纹理、形状和上下文等信息进行分类,提取出破坏建筑物和未破坏建筑物,试验表明,面向对象的分类方法应用于建筑物震害信息提取较传统分类方法有更高的精度,具有很大的应用潜力。
With the rapid development of remote sensing technology,especially more and more launches and applications of high resolution satellite,the technology plays an important role in identification of buildings,investigation and pre-evaluation of earthuake disaster.However,the traditional pixel-oriented approaches show a lot of limitations on classification of high resolution satellite images,and they can not take full advantages of image information.According to the characteristics of high-resolution remote sensing images,this paper presents a new object-oriented scheme of damaged building information extraction from high-resolution remote sensing image.The scheme has two steps.The first step is segmenting the whole imagery into image objects which do not intersect mutually.The second step is extracting damaged buildings and undamaged ones with the features used to classify,like spectrum,texture,shape and context.The experimental results indicate that the new object-oriented information extraction technique has high accuracy compared to the traditional classification methods and has a great application potential.
With the rapid development of remote sensing technology,especially more and more launches and applications of high resolution satellite,the technology plays an important role in identification of buildings,investigation and pre-evaluation of earthuake disaster.However,the traditional pixel-oriented approaches show a lot of limitations on classification of high resolution satellite images,and they can not take full advantages of image information.According to the characteristics of high-resolution remote sensing images,this paper presents a new object-oriented scheme of damaged building information extraction from high-resolution remote sensing image.The scheme has two steps.The first step is segmenting the whole imagery into image objects which do not intersect mutually.The second step is extracting damaged buildings and undamaged ones with the features used to classify,like spectrum,texture,shape and context.The experimental results indicate that the new object-oriented information extraction technique has high accuracy compared to the traditional classification methods and has a great application potential.
引文
[1]尹之潜.地震灾害及损失预测方法[M].北京:地震出版社.1995.
[2]朱博勤,魏成阶,张渊智.航空遥感地震灾害信息的快速提取[J].自然灾害学报.1998,7(1):34-38.
[3]柳稼航,单新建,尹京苑.遥感图像区域结构和纹理统计特性相结合的城市震害房屋自动识别[J].地震学报.2004,26(6):623-633.
[4]危福泉,姚新,陈琳.QuickBird数据在防震减灾管理信息系统中的应用[J].大地测量与地球动力学.2003.23(4):115-119.
[5]陈秋晓,骆剑承,周成虎,郑江,鲁学军,沈占锋.基于多特征的遥感影像分类方法[J].遥感学报.2004,8(3):239-245
[6]章毓晋.图像分割[M],北京:科学出版社,2001.
[2]朱博勤,魏成阶,张渊智.航空遥感地震灾害信息的快速提取[J].自然灾害学报.1998,7(1):34-38.
[3]柳稼航,单新建,尹京苑.遥感图像区域结构和纹理统计特性相结合的城市震害房屋自动识别[J].地震学报.2004,26(6):623-633.
[4]危福泉,姚新,陈琳.QuickBird数据在防震减灾管理信息系统中的应用[J].大地测量与地球动力学.2003.23(4):115-119.
[5]陈秋晓,骆剑承,周成虎,郑江,鲁学军,沈占锋.基于多特征的遥感影像分类方法[J].遥感学报.2004,8(3):239-245
[6]章毓晋.图像分割[M],北京:科学出版社,2001.
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心