Visual SLAM improves tracking and mapping with linear HDR images
RawSLAM: Online HDR Gaussian SLAM from Linear Radiance
Computer Vision and Pattern Recognition
Summary
Many current systems for mapping and tracking in visual environments struggle under very dark or very bright lighting because they use images with limited brightness information. The authors present a system that can work directly with very detailed, high dynamic range (HDR) images that capture a wider range of brightness, making the tracking and mapping more accurate in tough lighting. Their approach uses special mathematical methods to handle color and brightness efficiently, and it can also work on regular lower-bit images. They back up their method with a new dataset of indoor scenes and show improvements over previous methods.
What this means in practice
- •For robotics developers: Improve robot localization and mapping accuracy in challenging lighting conditions using linear HDR image inputs directly in SLAM pipelines.
- •For augmented reality engineers: Enhance AR device tracking stability and environmental reconstruction indoors under complex lighting by integrating the HDR Gaussian SLAM approach.
Authors
Marina Orozco González, Luis Merino
Abstract
Current dense visual SLAM systems rely almost exclusively on 8-bit tonemapped Low Dynamic Range (LDR) inputs, limiting their robustness in extreme lighting where shadows and highlights trigger tracking drift and mapping collapse. Conversely, existing raw and High Dynamic Range (HDR) reconstruction pipelines operate strictly offline. They depend on Structure-from-Motion preprocessing and are not suited for large inter-frame motion. We present, to the best of our knowledge, the first online Gaussian SLAM framework that tracks and maps directly on single-exposure 16-bit linear HDR imagery. Our method rests on three core components: an architecture-agnostic HDR Gaussian Splatting module featuring an MLP-free logarithmic parameterization of Gaussian color features; a Reinhard range-compressed photometric objective; and structure-guided spatial gradient weighting. Combined, these components allow our approach to outperform a direct HDR adaptation of MonoGS in both trajectory and reconstruction accuracy, while rendering natively in linear scene radiance for post-rendering processing. The same formulation runs unchanged on standard 8-bit inputs, roughly halving the MonoGS baseline error. Furthermore, our HDR Gaussian module transfers seamlessly to SplaTAM, Gaussian SLAM, and DROID-W, eliminating all tracking failures these systems suffer on challenging illumination sequences. To enable this research, we introduce RawSLAM: a dataset of 10 real-world indoor sequences featuring 16-bit RAW imagery, aligned depth, IMU measurements, and external OptiTrack poses. Code and dataset will be made publicly available soon.