Method of utilizing a two-dimensional image for detecting the position, posture, and shape of a three-dimensional objective
AssigneeMatsushita Electric Works, Ltd.
InventorsKatsuhiro Sasada, Tomoharu Nakahara, Hidekazu Araki
Life of this patent
Estimating the term…
Abstract
A method for recognizing a three-dimensional objective in which a two-dimensional image is obtained from the same viewing point as a distance image obtained by picking up the objective in a three-dimensional space and the two-dimensional image is utilized to previously limit an existence zone of the objective under detection for the distance image and to perform objective detecting operation over the partial distance image in the limited zone, thereby realizing a sufficiently high speed detection of, in particular, the three-dimensional objective.
Claims8
A method for detecting a position, posture and shape of an objective, the method comprising the steps of:
- obtaining a two-dimensional image of the objective from the same viewing point as a distance image obtained by photographing the objective in a three-dimensional space;
- specifying a partial image in the two-dimensional image;
- extracting a partial image of a distance image at a position corresponding to the position of the specified partial image of the two-dimensional image;
- and
- subjecting the partial image of the distance image to an objective detecting operation.
An objective recognition method comprising the steps of:
- obtaining a two-dimensional image from the same viewing point as a distance image obtained by photographing an objective in a three-dimensional space;
- specifying a partial image for the two-dimensional image;
- applying the specified two-dimensional partial image to the distance image to extract a partial image for the distance image;
- and
- subjecting the distance partial image to an objective detecting operation, wherein three-dimensional information obtained from the distance partial image is applied to a plane, the three-dimensional information is rotated on an origin of a camera coordinate system to cause a normal of the plane to coincide with a Z axis of the camera coordinate system, the three-dimensional information is then projected onto an X-Y plane of the coordinate system, a projected image is collated with a previously-registered model to find a temporary position and posture of the target objective, the temporary position and posture are subjected to an inverse transformation to calculate a position and posture of the target objective before the rotation.
Dependencies are inferred from the claim text.