DOI: 10.1016/j.jag.2013.05.017
Scopus记录号: 2-s2.0-84897585667
论文题名: A framework for mapping tree species combining hyperspectral and LiDAR data: Role of selected classifiers and sensor across three spatial scales
作者: Ghosh A ; , Fassnacht F ; E ; , Joshi P ; K ; , Kochb B
刊名: International Journal of Applied Earth Observation and Geoinformation
ISSN: 15698432
出版年: 2014
卷: 26, 期: 1 起始页码: 49
结束页码: 63
语种: 英语
英文关键词: Hyperspectral
; nDSM
; Random forest (RF)
; Scale
; Support vector machines (SVM)
; Tree species classificationa
Scopus关键词: image analysis
; image classification
; lidar
; mapping
; pixel
; random walk method
; remote sensing
; satellite sensor
; sensor
; spatial analysis
; spatial data
; spatial resolution
; spectral analysis
; tree
; vegetation index
; Europe
英文摘要: Knowledge of tree species distribution is important worldwide for sustainable forest management andresource evaluation. The accuracy and information content of species maps produced using remotesensing images vary with scale, sensor (optical, microwave, LiDAR), classification algorithm, verificationdesign and natural conditions like tree age, forest structure and density. Imaging spectroscopy reducesthe inaccuracies making use of the detailed spectral response. However, the scale effect still has a stronginfluence and cannot be neglected. This study aims to bridge the knowledge gap in understanding thescale effect in imaging spectroscopy when moving from 4 to 30 m pixel size for tree species mapping,keeping in mind that most current and future hyperspectral satellite based sensors work with spatialresolution around 30 m or more.Two airborne (HyMAP) and one spaceborne (Hyperion) imaging spectroscopy dataset with pixel sizesof 4, 8 and 30 m, respectively were available to examine the effect of scale over a central European forest.The forest under examination is a typical managed forest with relatively homogenous stands featuringmostly two canopy layers. Normalized digital surface model (nDSM) derived from LiDAR data was usedadditionally to examine the effect of height information in tree species mapping. Six different sets ofpredictor variables (reflectance value of all bands, selected components of a Minimum Noise Fraction(MNF), Vegetation Indices (VI) and each of these sets combined with LiDAR derived height) were exploredat each scale. Supervised kernel based (Support Vector Machines) and ensemble based (Random Forest)machine learning algorithms were applied on the dataset to investigate the effect of the classifier. Iterativebootstrap-validation with 100 iterations was performed for classification model building and testing forall the trials.For scale, analysis of overall classification accuracy and kappa values indicated that 8 m spatial res-olution (reaching kappa values of over 0.83) slightly outperformed the results obtained from 4 m forthe study area and five tree species under examination. The 30 m resolution Hyperion image producedsound results (kappa values of over 0.70), which in some areas of the test site were comparable withthe higher spatial resolution imagery when qualitatively assessing the map outputs. Considering inputpredictor sets, MNF bands performed best at 4 and 8 m resolution. Optical bands were found to be best for30 m spatial resolution. Classification with MNF as input predictors produced better visual appearanceof tree species patches when compared with reference maps. Based on the analysis, it was concludedthat there is no significant effect of height information on tree species classification accuracies for thepresent framework and study area. Furthermore, in the examined cases there was no single best choiceamong the two classifiers across scales and predictors. It can be concluded that tree species mappingfrom imaging spectroscopy for forest sites comparable to the one under investigation is possible withreliable accuracies not only from airborne but also from spaceborne imaging spectroscopy datasets. © 2013 Elsevier B.V. Citation statistics:
资源类型: 期刊论文
标识符: http://119.78.100.158/handle/2HF3EXSE/79802
Appears in Collections: 气候变化事实与影响
There are no files associated with this item.
作者单位: Department of Natural Resources, TERI University, Vasant Kunj, New Delhi, India; Remote Sensing and Landscape Information Systems, University of Freiburg, Tennenbacherstraße. 4, D-79106 Freiburg, Germany
Recommended Citation:
Ghosh A,, Fassnacht F,E,et al. A framework for mapping tree species combining hyperspectral and LiDAR data: Role of selected classifiers and sensor across three spatial scales[J]. International Journal of Applied Earth Observation and Geoinformation,2014-01-01,26(1)