Camera-based techniques provide centimeter-level, 6 Degree-of-Freedom positioning that is well beyond the accuracy of other methods. As it is implemented in the GoodMaps Explore app today, the user’s smartphone camera captures images which are analyzed for unique points to be compared to features in the images collected in the initial building mapping scan. Then using Artificial Intelligence/machine learning, it can reverse lookup the position based on those unique points. This method provides positioning accuracy up to 2-3 ft.
How accurate is Camera-based positioning?