Benchmarking SLAM Performance in Low-Light Environments
July 19, 2026
This report benchmarks six SLAM systems, including ORB-SLAM3, DSO, and DPVO, across five LaMARia sequences with varying illumination levels. The study evaluates how reduced contrast and sensor noise impact feature extraction in standard RGB camera-based robotics.
HOW THIS AFFECTS YOU
●
builderYou can select more robust SLAM frameworks for autonomous robots operating in dark or poorly lit settings.
●
researcherYou can use these benchmarks to evaluate learning-based versus classical feature-matching paradigms in low-light conditions.