Abstract:
The accuracy of hotspot signatures is crucial to the development of algorithms for the retrieval of various biophysical parameters of terrestrial surface targets. The hot...View moreMetadata
Abstract:
The accuracy of hotspot signatures is crucial to the development of algorithms for the retrieval of various biophysical parameters of terrestrial surface targets. The hotspot effect determined by the kernel-driven RossThick-LiSparseReciprocal (RTLSR) Bidirectional Reflectance Distribution Function (BRDF) model fully relies on the performance of the two kernels in modeling the hotspot effect. Previously, a method has been developed to correct the volumetric scattering component of the RTLSR model (Jiao et al., 2016); however, in few cases that the weight of the volumetric scattering component (f
vol
) is no longer significant (e.g., f
vol
= 0) in the framework of the RTLSR model, the slight underestimation of the hotspot effect still exists. In this study, we propose a method to enhance the overlap function inherent in the geometric-optical (GO) kernel using a physical hotspot factor. The hotspot observations extracted from the entire archive of the POLDER-3 BRDF database are used to determine two parameters of this hotspot factor. Result shows that the proposed method further improves the model-observation fits in the vicinity of hotspot direction, particularly in some extreme cases where the GO scattering component is fully dominant in the multiangle measurements. Such an improved GO kernel, combining with the hotspot adjustment method for the volumetric scattering kernel, reconstructs hotspot effect more accurately; therefore, necessarily further improves the performance of the kernel-driven BRDF model particularly for the retrieval of the canopy structure parameters that is associated with the hotspot effect.
Date of Conference: 22-27 July 2018
Date Added to IEEE Xplore: 04 November 2018
ISBN Information: