期刊名称：ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
印刷版ISSN：2194-9042
电子版ISSN：2194-9050
出版年度：2010
卷号：XXXVIII Part 4
出版社：Copernicus Publications
摘要：Spatially heterogeneous patterns of land use in urban environments have long posed a challenge to remote sensing. High spatial resolutionpassive sensors provide detailed data of urban regions at sub meter level but are frequently limited by shadows of the built environment.Moderate resolution data can provide synoptic perspectives of such landscapes but tend to obscure information of spectrally similarobjects. Due to its height-above-ground component, which is unaffected by shadows, Light Detection and Ranging (LiDAR) data areincreasingly being used as an alternative to passive sensors. However, LiDAR’s intensity component is infrequently utilized in urbanstudies presumably because its range of digital number values is similar between urban impervious and tree canopy covers. Previousinvestigations have concentrated on mapping either tree canopy or buildings using local-scale normalization procedures but the use ofnormalized intensity to map multiple land-use types in a heterogeneous urban landscape at a regional scale has received little attention.Our approach uniquely utilizes normalized intensity data in combination with structural components derived from LiDAR masspointsusing maximum likelihood estimation of land use classes. Preliminary results show that our approach accurately distinguishes impervioussurfaces and tree canopy over broad metropolitan contexts, with an overall accuracy of 96.7% for the ML classification of integratedLiDAR. In summary, we found that normalized LiDAR intensity data can be integrated with LiDAR surface models improving ourability to map heterogeneous urban geographies.
关键词：LiDAR data; intensity; normalization; high resolution; urban landscape; heterogeneity
