Supercharging Thermal Gaussian Splatting with Depth Estimation


Manoj Biswanath*1, Chenxin Cai*1, Hannah Schieber2, Daniel Roth2, Benjmanin Busam1

* denotes equal contribution


Photogrammetry and Remote Sensing, Munich Center for Machine Learning (MCML) Technical University of Munich1
Human-Centered Computing and Extended Reality Lab, TUM University Hospital, Clinic for Orthopedics and Sports Orthopedics, Munich Institute of Robotics and Machine Intelligence (MIRMI)2

Arxiv — Coming Soon
accepted ISPRS congress 2026
Code

Abstract

Efficient and robust 3D scene representation is crucial in fields such as robotics, autonomous driving, and augmented reality. While RGB images provide valuable content for 3D reconstruction, other modalities like thermal or depth can enable additional information on the 3D environment. Lately, novel view synthesis methods like 3D Gaussian Splatting have started using multiple modalities to further boost their performance. But fusing or combining those multi-modal data can make the process slower and can bring in additional challenges. Therefore, our project aims to use single modality based on thermal infrared domain, by removing the reliance on visible light, as much as possible. This single modality can be expected to be faster as it doesn’t rely on multimodal data. We propose a method Thermal-to-Depth Gaussian (TDg), that uses only thermal images and depth estimation in its architecture to derive the thermal radiance fields. Our method performs better than the MSMG (Multiple Single-Modal Gaussians) baseline in most cases on our test datasets, RGBT-Scenes and ThermalMix. On average, the rendering quality metrics PSNR, SSIM, and LPIPS of our TDg method are 27.42, 0.8899, 0.175, which are better than the baseline MSMG values 27.18, 0.8896, 0.177. It also reduces the training time significantly, from 0.389h to 0.176h. Overall our method is successful in deriving these 3D thermal models/radiance fields, which can ultimately have several applications, such as identifying heat sources critical in surveillance, search and rescue operations, industrial inspections where temperature is widely used to monitor machines.



Citation

                @misc{tdg2026,
                  title        = {Supercharging Thermal Gaussian Splatting with Depth Estimation},
                  author       = {Biswanath, Manoj, and Cai, Chenxin, and Schieber, Hannah, and Roth, Daniel, and Benjamin, Busam},
                  year         = {2026},
                }