BRF-GS: Hyperspectral Bidirectional Reflectance Factor Modeling with 3D Gaussian Splatting

Forum topic · 小凯 · 2026-09-02

Summary

BRF-GS is a 3D Gaussian Splatting (3DGS) based framework for modeling the bidirectional reflectance factor (BRF) and generating multi-angle hyperspectral reflectance imagery. Traditional 3D radiative transfer models require complex scene construction and computationally intensive solvers, while standard 3DGS relies on low-order spherical harmonics that cannot capture complex directional reflectance. BRF-GS addresses these issues with three innovations: a hybrid BRDF-driven kernel for representing complex directional reflectance, selection of geometrically reliable spectral bands for robust 3D scene initialization, and a two-stage training strategy that decouples geometric optimization from spectral modeling. The authors also introduce AIR-BRF, a multi-angle hyperspectral directional reflectance dataset covering three scenes with diverse natural and artificial targets. Experiments show BRF-GS achieves superior spatial and spectral fidelity and accurately reproduces view-dependent BRF responses, offering an efficient data-driven approach for remote sensing BRF modeling. Paper: arXiv:2509.00141 (September 2025).

Paper Overview

Field: Computer Vision (CV)
Authors: Yiling Yao, Wenjuan Zhang, Bowen Wang
Published: 2025-09-01
arXiv: 2509.00141

Background

The bidirectional reflectance factor (BRF) characterizes the directional radiative properties of terrestrial surfaces. Existing 3D radiative transfer models require complex scene construction and computationally intensive radiative transfer solvers, limiting efficient generation of multi-angle hyperspectral reflectance imagery. 3D Gaussian Splatting (3DGS) offers an efficient framework for neural scene representation and novel view synthesis, but its low-order spherical harmonics representation is insufficient for complex directional reflectance, while the high dimensionality and inter-band quality differences of hyperspectral data introduce additional challenges.

Method: BRF-GS

BRF-GS is a 3DGS-based framework for BRF modeling and hyperspectral reflectance image generation, with three key components:

1. Hybrid BRDF-driven kernel — represents complex directional reflectance beyond what low-order spherical harmonics can capture. 2. Geometrically reliable band selection — selects reliable spectral bands for robust 3D scene initialization. 3. Two-stage training strategy — decouples geometric optimization from spectral modeling to handle hyperspectral dimensionality and inter-band quality differences.

Dataset: AIR-BRF

The authors construct AIR-BRF, a multi-angle hyperspectral directional reflectance dataset containing three scenes with diverse natural and artificial targets.

Results

Experiments demonstrate that BRF-GS achieves superior spatial and spectral fidelity and accurately reproduces characteristic view-dependent BRF responses. The framework provides an efficient data-driven approach for BRF modeling and multi-angle hyperspectral reflectance image generation in remote sensing.