DOI: 10.1016/j.jag.2014.08.017
Scopus记录号: 2-s2.0-84924367249
论文题名: Automated metric characterization of urban structure using buildingdecomposition from very high resolution imagery
作者: Heinzel J ; , Kemper T
刊名: International Journal of Applied Earth Observation and Geoinformation
ISSN: 15698432
出版年: 2015
卷: 35, 期: PB 起始页码: 151
结束页码: 160
语种: 英语
英文关键词: Built-up metrics
; Mathematical morphology
; South Africa
; Urban characterization
; Very high resolution imagery
Scopus关键词: accuracy assessment
; aerial photograph
; aerial survey
; algorithm
; building
; decomposition analysis
; image classification
; image resolution
; urban area
; urban development
; North West Province
; Rustenburg
; South Africa
英文摘要: Classification approaches for urban areas are mostly of qualitative and semantic nature. They produceinterpreted classes similar to those from land cover and land use classifications. As a complement to thoseclasses, quantitative measures directly derived from the image could lead to a metric characterization ofthe urban area. While these metrics lack of qualitative interpretation they are able to provide objectivemeasure of the urban structures. Such quantitative measures are especially important in rapidly growing cities since, beside of thegrowth in area, they can provide structural information for specific areas and detect changes. Rusten-burg, which serves as test area for the present study, is amongst the fastest growing cities in SouthAfrica. It reveals a heterogeneous face of housing and building structures reflecting social and/or eco-nomic differences often linked to the spatial distribution of industrial and local mining sites. Up to datecoverage with aerial photographs is provided by aerial surveys in regular intervals. Also recent satellitesystems provide imagery with suitable resolution. Using such set of very high resolution images a fullyautomated algorithm has been developed which outputs metric classes by systematically combiningimportant measures of building structure. The measurements are gained by decomposition of buildingsdirectly from the imagery and by using methods from mathematical morphology. The decomposed build-ing objects serve as basis for the computation of grid statistics. Finally a systematic combination of thesingle features leads to combined metrical classes. For the dominant urban structures verification results indicate an overall accuracy of at least 80% onthe single feature level and 70% for the combined classes. © 2014 Elsevier B.V. Citation statistics:
资源类型: 期刊论文
标识符: http://119.78.100.158/handle/2HF3EXSE/79478
Appears in Collections: 气候变化事实与影响
There are no files associated with this item.
作者单位: Joint Research Centre of the European Commission, Institute for the Protection and Security of the Citizen, Global Security and Crisis Management Unit, Via E. Fermi 2749, Ispra (VA), Italy
Recommended Citation:
Heinzel J,, Kemper T. Automated metric characterization of urban structure using buildingdecomposition from very high resolution imagery[J]. International Journal of Applied Earth Observation and Geoinformation,2015-01-01,35(PB)