No.4（2009）10.Environmental map generation applying scale invariant feature transform algorithm

Norihiro Ohira, Shinichi Tomiyama

  A method of environmental map generation is reported herein. In the case of generating an environmental map, we usually input featureless images like images of the ceiling or a floor and their illumination change. The SIFT (Scale Invariant Feature Transform) algorithm detects local feature points and extracts image features that are reasonably invariant to changes in illumination, image noise, rotation, scaling, and small changes in viewpoint.
  In this paper, we focused our attention on the image composition of featureless images applying the SIFT algorithm and proposed one simple method. The proposed method could run in the order of n log n time to compose two images of order n pixels. Also, the image composition was precise enough to be used, inputting featureless data like images of a ceiling.

Keywords

Environmental map, SIFT algorithm, images compose

Back to Contents