OctoMap is an open-source framework for generating volumetric 3D environment models, based on octrees and probabilistic occupancy estimation. It explicitly represents occupied, free, and unknown areas, and includes a compression method to keep the 3D models compact. The framework has been successfully applied in various robotics projects and has been evaluated using real-world datasets. The results demonstrate efficient and consistent updates, minimal memory requirements, and the ability to model complex environments. The framework is available as a C++ library under the BSD license and can be integrated into robotics systems, including the Robot Operating System (ROS). The paper discusses related work, implementation details, and experimental results, highlighting the framework's advantages in terms of probabilistic representation, efficiency, and compactness.OctoMap is an open-source framework for generating volumetric 3D environment models, based on octrees and probabilistic occupancy estimation. It explicitly represents occupied, free, and unknown areas, and includes a compression method to keep the 3D models compact. The framework has been successfully applied in various robotics projects and has been evaluated using real-world datasets. The results demonstrate efficient and consistent updates, minimal memory requirements, and the ability to model complex environments. The framework is available as a C++ library under the BSD license and can be integrated into robotics systems, including the Robot Operating System (ROS). The paper discusses related work, implementation details, and experimental results, highlighting the framework's advantages in terms of probabilistic representation, efficiency, and compactness.