An Approach for Validating Roundabout Using Spline and Curvature for Map Data

A 2D Map is a collection of pure lines connected to form unique features like roundabouts, cul-de-sacs, buildings, other figures etc. In practical maps created, lines using latitude and longitude. The idea in this paper identifies and validates the roundabouts utilizing geometry of the Map features by spline technique where the geometries are passed to the nonrational uniform B spline function, and the new control points enabled us to calculate the curvature (κ) and radius. We constructed the bounding box of the roundabout geometry and then carved the best-fitted curve, which can be either of eclipse or a circle. We considered the average of two radii in the case of an eclipse. Finally, the radius is bounded between the threshold and checked with the reciprocal of the calculated curvature from the spline. With a threshold of 10 m, around 92.2% of roundabouts in Europe were validated. Improvement of approx. 12.2% in the deterministic technique have been achieved regarding the prior art, which detects the roundabout in a deterministic way to 80%.