Object-Based Shadow Index via Illumination Intensity from High Resolution Satellite Images over Urban Areas

For multi-spectral remote sensing imagery, accurate shadow extraction is of great significance for overcoming the information loss caused by high buildings and the solar incidence angle in urban remote sensing. However, diverse solar illumination conditions, similarities between shadows, and other d...

Full description

Bibliographic Details
Main Authors: Haoyang Fu, Tingting Zhou, Chenglin Sun
Format: Article
Language:English
Published: MDPI AG 2020-02-01
Series:Sensors
Subjects:
Online Access:https://www.mdpi.com/1424-8220/20/4/1077
_version_ 1828116979180896256
author Haoyang Fu
Tingting Zhou
Chenglin Sun
author_facet Haoyang Fu
Tingting Zhou
Chenglin Sun
author_sort Haoyang Fu
collection DOAJ
description For multi-spectral remote sensing imagery, accurate shadow extraction is of great significance for overcoming the information loss caused by high buildings and the solar incidence angle in urban remote sensing. However, diverse solar illumination conditions, similarities between shadows, and other dark land features bring uncertainties and deviations to shadow extraction processes and results. In this paper, we classify shadows as either strong or weak based on the ratio between ambient light intensity and direct light intensity, and use the fractal net evolution approach (FNEA), which is a multi-scale segmentation method based on spectral and shape heterogeneity, to reduce the interference of salt and pepper noise and relieve the error of misdiagnosing land covers with high reflectivity in shaded regions as unshaded ones. Subsequently, an object-based shadow index (OSI) is presented according to the illumination intensities of different reflectance features, as well as using the normalized difference water index (NDWI) and near infrared (NIR) band to highlight shadows and eliminate water body interference. The data from three high-spatial-resolution satellites—WorldView-2 (WV-2), WorldView-3 (WV-3), and GaoFen-2 (GF-2)—were used to test the methods and verify the robustness of the OSI. The results show that the OSI index performed well regarding both strong and weak shadows with the user accuracy and the producer accuracy both above 90%, while the four other existing indexes that were tested were not effective at diverse solar illumination conditions. In addition, all the disturbances from water body were excluded well when using the OSI, except for the GF-2 data in weak shadows.
first_indexed 2024-04-11T13:05:37Z
format Article
id doaj.art-155472467d2d4f8d84ea31612eeb0773
institution Directory Open Access Journal
issn 1424-8220
language English
last_indexed 2024-04-11T13:05:37Z
publishDate 2020-02-01
publisher MDPI AG
record_format Article
series Sensors
spelling doaj.art-155472467d2d4f8d84ea31612eeb07732022-12-22T04:22:45ZengMDPI AGSensors1424-82202020-02-01204107710.3390/s20041077s20041077Object-Based Shadow Index via Illumination Intensity from High Resolution Satellite Images over Urban AreasHaoyang Fu0Tingting Zhou1Chenglin Sun2Coherent Light and Atomic and Molecular Spectroscopy Laboratory, College of Physics, Jilin University, Changchun 130012, ChinaCoherent Light and Atomic and Molecular Spectroscopy Laboratory, College of Physics, Jilin University, Changchun 130012, ChinaCoherent Light and Atomic and Molecular Spectroscopy Laboratory, College of Physics, Jilin University, Changchun 130012, ChinaFor multi-spectral remote sensing imagery, accurate shadow extraction is of great significance for overcoming the information loss caused by high buildings and the solar incidence angle in urban remote sensing. However, diverse solar illumination conditions, similarities between shadows, and other dark land features bring uncertainties and deviations to shadow extraction processes and results. In this paper, we classify shadows as either strong or weak based on the ratio between ambient light intensity and direct light intensity, and use the fractal net evolution approach (FNEA), which is a multi-scale segmentation method based on spectral and shape heterogeneity, to reduce the interference of salt and pepper noise and relieve the error of misdiagnosing land covers with high reflectivity in shaded regions as unshaded ones. Subsequently, an object-based shadow index (OSI) is presented according to the illumination intensities of different reflectance features, as well as using the normalized difference water index (NDWI) and near infrared (NIR) band to highlight shadows and eliminate water body interference. The data from three high-spatial-resolution satellites—WorldView-2 (WV-2), WorldView-3 (WV-3), and GaoFen-2 (GF-2)—were used to test the methods and verify the robustness of the OSI. The results show that the OSI index performed well regarding both strong and weak shadows with the user accuracy and the producer accuracy both above 90%, while the four other existing indexes that were tested were not effective at diverse solar illumination conditions. In addition, all the disturbances from water body were excluded well when using the OSI, except for the GF-2 data in weak shadows.https://www.mdpi.com/1424-8220/20/4/1077shadow indexmulti-scale segmentationobject-basedillumination intensityshadow intensityurban remote sensing
spellingShingle Haoyang Fu
Tingting Zhou
Chenglin Sun
Object-Based Shadow Index via Illumination Intensity from High Resolution Satellite Images over Urban Areas
Sensors
shadow index
multi-scale segmentation
object-based
illumination intensity
shadow intensity
urban remote sensing
title Object-Based Shadow Index via Illumination Intensity from High Resolution Satellite Images over Urban Areas
title_full Object-Based Shadow Index via Illumination Intensity from High Resolution Satellite Images over Urban Areas
title_fullStr Object-Based Shadow Index via Illumination Intensity from High Resolution Satellite Images over Urban Areas
title_full_unstemmed Object-Based Shadow Index via Illumination Intensity from High Resolution Satellite Images over Urban Areas
title_short Object-Based Shadow Index via Illumination Intensity from High Resolution Satellite Images over Urban Areas
title_sort object based shadow index via illumination intensity from high resolution satellite images over urban areas
topic shadow index
multi-scale segmentation
object-based
illumination intensity
shadow intensity
urban remote sensing
url https://www.mdpi.com/1424-8220/20/4/1077
work_keys_str_mv AT haoyangfu objectbasedshadowindexviailluminationintensityfromhighresolutionsatelliteimagesoverurbanareas
AT tingtingzhou objectbasedshadowindexviailluminationintensityfromhighresolutionsatelliteimagesoverurbanareas
AT chenglinsun objectbasedshadowindexviailluminationintensityfromhighresolutionsatelliteimagesoverurbanareas