Sidewalk Detection

Sidewalk detection research focuses on automatically identifying and mapping sidewalks from various data sources, primarily to improve urban planning, accessibility assessments, and public safety. Current approaches utilize a range of techniques, including ensemble learning models that outperform large language models in robustness to noisy conditions, and Structure from Motion (SfM) methods to locate sidewalks even when obscured. These advancements enable the creation of detailed, accurate sidewalk maps and the identification of hazards, ultimately contributing to more efficient urban infrastructure management and improved accessibility for pedestrians.