Mapping in autonomous and non-autonomous platforms

This invention describes a way for robots and autonomous vehicles to know exactly where they are by creating and constantly improving a 3D map of their surroundings. It uses cameras and motion sensors to gather data, then refines the map by checking how the 3D points in the map align with what the camera sees in 2D images, correcting any errors to make the machine's location awareness faster and more accurate. The claims specifically focus on methods for correcting and refining these 3D maps.