VSLAM with Image Enhancement Method LIME for Supporting Rescue in Dark Areas

Rescue operations in hazardous environments, such as collapsed buildings or unexplored caves, require effective tools to navigate complex and obstructive terrains. Visual Simultaneous Localization and Mapping (VSLAM) is a valuable system in these situations, as it helps generate maps of unknown areas, guiding rescuers to victims while minimizing risk. However, low-light conditions can impair VSLAM's accuracy. While existing methods like EnlightenGAN have been used to enhance low-light images, they are resource-intensive. This study proposes LIME (Low-light Image Enhancement via Illumination Map Estimation) as a more cost-effective alternative, requiring less computational power and no collection of training dataset while maintaining competitive performance. Testing with the ORB-SLAM3 system shows that improved LIME outperforms EnlightenGAN in dark conditions, offering a promising solution for rescue operations.