This paper presents a method based on 3D invariant features for the recognition of three-dimensional object. In this
method, the projective invariant features are adopted as the representative of the 3D invariance. 3D invariant features are
the features that will not change with viewpoints or poses, so they can be used to recognize three-dimensional objects.
Exact point-line matching is achieved in use of the point-line characteristics obtained. On this basis, the views relation is
established by the 2D projective transformation, and many kinds of 3D invariant features of complicated spatial
structures can be neatly extracted by combining real elements and virtual elements. Therefore the three-dimensional
object can be recognized. The results indicate that the method could greatly satisfy three-dimensional object recognition.