3D models are an essential part of many robotic applications. In applications\nwhere the environment is unknown a-priori, or where only a part of the\nenvironment is known, it is important that the 3D model can handle the unknown\nspace efficiently. Path planning, exploration, and reconstruction all fall into\nthis category. In this paper we present an extension to OctoMap which we call\nUFOMap. UFOMap uses an explicit representation of all three states in the map,\ni.e., occupied, free, and unknown. This gives, surprisingly, a more memory\nefficient representation. Furthermore, we provide methods that allow for\nsignificantly faster insertions into the octree. This enables real-time colored\nvolumetric mapping at high resolution (below 1 cm). UFOMap is contributed as a\nC++ library that can be used standalone but is also integrated into ROS.\n