作者：
Luo, HuiLi, DerenLiu, ChongWuhan Univ
State Key Lab Informat Engn Surveying Mapping & R Wuhan Hubei Peoples R China Wuhan Univ
Collaborat Innovat Ctr Geospatial Technol Wuhan Hubei Peoples R China Jiangxi Normal Univ
Key Lab Poyang Lake Wetland & Watershed Res Minist Educ Nanchang Jiangxi Peoples R China
object-based shadow detection in urban areas is an important topic in very high resolution remote sensing image processing. Multi-resolution segmentation (MRS) is an effective segmentation method, and is used for obje...
详细信息
object-based shadow detection in urban areas is an important topic in very high resolution remote sensing image processing. Multi-resolution segmentation (MRS) is an effective segmentation method, and is used for object-based shadow detection. However, several input parameters within MRS may result in unstable performance for final shadowdetection;thus, the evaluation and optimization for the parameters upon the final shadowdetection accuracy cannot be overlooked. In this paper, the three parameters in MRS (scale s, weight of colour w(color) and weight of compactness w(compact)) upon the final result of a recently proposed method, object-based shadow detection with Dempster-Shafer theory, were evaluated and optimized by sensitivity analysis and Taguchi's method with three experimental data. Experiments show that scale s is the most sensitive parameter among the three parameters within MRS. More importantly, according to the Taguchi's method theory, there is a very significant interaction effect between s and w(color), which cannot be overlooked. The shadowdetection accuracy yielded by the optimum parameter combination in consideration of the interaction effect is higher than that only optimized by covering the main effect of single parameter in most cases.
shadows commonly exist in high resolution satellite imagery, particularly in urban areas, which is a combined effect of low sun elevation, off-nadir viewing angle, and high-rise buildings. The presence of shadows can ...
详细信息
shadows commonly exist in high resolution satellite imagery, particularly in urban areas, which is a combined effect of low sun elevation, off-nadir viewing angle, and high-rise buildings. The presence of shadows can negatively affect image processing, including land cover classification, mapping, and object recognition due to the reduction or even total loss of spectral information in shadows. The compensation of spectral information in shadows is thus one of the most important preprocessing steps for the interpretation and exploitation of high resolution satellite imagery in urban areas. In this study, we propose a new approach for global shadow compensation through the utilization of fully constrained linear spectral unmixing. The basic assumption of the proposed method is that the construction of the spectral scatter plot in shadows is analogues to that in non-shadow areas within a two-dimension spectral mixing space. In order to ensure the continuity of land covers, a smooth operator is further used to refine the restored shadow pixels on the edge of non-shadow and shadow areas. The proposed method is validated using the WorldView-2 multispectral imagery collected from downtown Toronto, Ontario, Canada. In comparison with the existing linear-correlation correction method, the proposed method produced the compensated shadows with higher quality. (C) 2014 Elsevier B.V. All rights reserved.
暂无评论