Fast generation of an environment map for an indoor mobile robot by image reduction

A fast and efficient algorithm to build an environment map is presented for the stereo vision of indoor mobile robots, by using image reduction and parallel processing. The real images of corridors are complex images and may have weak contrast under natural lighting condition, so large reduction may spoil the accuracy of an environment map. To keep the accuracy of the map, the pyramid in this paper has only two levels. To reduce the number of computations minimally, its reduction scale can be any size including 1/2ⁿ. Also, the parallel processing is used for fast computation. A data structure of line support region and the fast merging algorithm are proposed. The proposed algorithm is compared with an algorithm that uses only original images. The experimental results show the proposed algorithm saves processing time, which takes only 7% of the conventional algorithm, and preserves the accuracy of an environment map