Semantic Mapping

Semantic mapping aims to create rich, environment representations that go beyond simple geometry, incorporating object identities and locations. Current research focuses on improving the accuracy and robustness of these maps, particularly in challenging environments, using techniques like Bayesian Kernel Inference, evidential reasoning (e.g., Dempster-Shafer theory), and neural implicit representations, often integrated with advanced object detection and segmentation methods (e.g., YOLO, vision-language models). These advancements are crucial for enabling autonomous navigation in robotics, improving autonomous driving safety, and facilitating applications in fields like agriculture and search and rescue.