Diffusion-Denoised Hyperspectral Gaussian Splatting

TL;DR

This paper introduces DD-HGS, a novel 3D hyperspectral scene reconstruction method that combines diffusion denoising with Gaussian splatting, achieving state-of-the-art results in spectral and spatial accuracy.

Contribution

It presents a new explicit 3D hyperspectral reconstruction approach integrating diffusion denoising, wavelength-aware spherical harmonics, and spectral loss, improving over existing methods.

Findings

Achieves state-of-the-art performance on Hyper-NeRF dataset.

Effectively reconstructs hyperspectral scenes with high spectral fidelity.

Demonstrates faster training and rendering compared to prior methods.

Abstract

Hyperspectral imaging (HSI) has been widely used in agricultural applications for non-destructive estimation of plant nutrient composition and precise quantification of sample nutritional elements. Recently, 3D reconstruction methods, such as Neural Radiance Field (NeRF), have been used to create implicit neural representations of HSI scenes. This capability enables the rendering of hyperspectral channel compositions at every spatial location, thereby helping localize the target object's nutrient composition both spatially and spectrally. However, it faces limitations in training time and rendering speed. In this paper, we propose Diffusion-Denoised Hyperspectral Gaussian Splatting (DD-HGS), which enhances the state-of-the-art 3D Gaussian Splatting (3DGS) method with wavelength-aware spherical harmonics, a Kullback-Leibler divergence-based spectral loss, and a diffusion-based denoiser to enable 3D explicit reconstruction of the hyperspectral scenes for the entire spectral range. We present extensive evaluations on diverse real-world hyperspectral scenes from the Hyper-NeRF dataset to show the effectiveness of our DD-HGS. The results demonstrate that DD-HGS achieves the new state-of-the-art performance compared to all the previously published methods. Project page: https://dragonpg2000.github.io/DDHGS-website/