Invariant feature extraction of curved patches similar to quadric with laser points cloud

Abstract

Feature extraction is the key technology in 3D modeling and recognition. By using points cloud data's processing methods based on grid and points set synthetically, first it was proposed that a watershed region increasing segmentation method based on binary morphology for range image, then it was researched that fitting parameter equation of quadric patch by adopting the eigenvector estimate method of linear least squares, finally it were analyzed that the extraction technologies of differential coefficient invariant and movement invariant under Euclid transform. © 2011 Springer-Verlag.