Back to Search Start Over

Study on directional reflectivity characteristics analysis of mixed pixels using multi-angle spectral measurements

Authors :
Junlong Xu
Xingping Wen
Dayou Luo
Haonan Zhang
Source :
Earth Science Informatics.
Publication Year :
2021
Publisher :
Springer Science and Business Media LLC, 2021.

Abstract

The influence of the detection direction on the quality of spectral data was not taken into account, which is catastrophic for some samples where the reflectivity direction was very different. To explore the directional characteristics of mixed pixels reflectivity, we used an ASD FieldSpec3 spectroradiometer to carry out multi-angle spectral measurement experiments in the laboratory, and the direction error model is established. As a result, the proportions of black area (PBA) and the observed azimuths (Φ) will affect the reflectivity of the mixed pixel in the visible light band. It is found that when the PBA is close to 0, the reflectivity distribution of mixed pixels is characterized by the reflected energy reflecting uniformly around the entire hemisphere space. When PBA is close to 1, there is a significant difference in the reflectivity of mixed pixels in 2π space. The direction error model better reflects the reflectivity changes caused by PBA and observed azimuth. The mean absolute error of the estimated reflectivity compared with the measured value is only 0.047. When the PBA is large, the estimation accuracy of the model is higher. When the PBA is small, and the observed azimuth is large, the accuracy of the model is slightly worse. The “two-block” mixed pixel is an ideal sample to satisfy the direction error model, while a dispersed “multi-block” mixed pixel is not applicable to the correction model. When conducting mixed pixel spectral measurement experiment, the influence of the detection azimuth on the spectral reflectivity should be fully considered, which is beneficial to improve the reliability of experiments on multi-angle spectral measurements of mixed pixels.

Details

ISSN :
18650481 and 18650473
Database :
OpenAIRE
Journal :
Earth Science Informatics
Accession number :
edsair.doi...........df8e6bc18fd8dfe66e265870ac548982
Full Text :
https://doi.org/10.1007/s12145-021-00622-5