Procedural Modeling of Terrain from GPS Routes

Procedural modeling techniques provide an easy way for realistic terrain synthesis. There are many methods for this purpose, but few of them provide the necessary constraints to control the final result. The current controlling methods are slow and inaccurate. The goal of this paper is to present a new procedural method to synthesis realistic terrain that fulfills the constraint of a set of consecutive coordinates from GPS routes. In the new method, random terrain generation is based on Perlin noise algorithm. Instead of getting a random value for all points of each octave, the method presents the novelty of solving the needed values of each point to fulfill the coordinates of the GPS route when Perlin algorithm is applied. Points that are not included in the constraint, keep the randomness provided by the original Perlin method. The results show that the method allows, at low computational cost, to integrate different complexity paths with high accuracy while conserving the natural aspect of the procedural generation.