Object-shape generation method, object-shape generation apparatus, and program

摘要

Even in a situation in which an image of the back side of an object cannot be captured, the shape of the object can be generated with high accuracy. For individual images captured from the circumference of an object, projection regions projected in an area from the surface of projection in a projective space to a projective plane that is distant from the surface of projection by a predetermined length in a depth direction in a case where the object reflected in the individual images is projected to the projective space from viewpoint positions of the individual images are detected. A portion common to the individual detected projection regions is extracted.

主权项

1. An object-shape generation method, comprising:
in an apparatus: capturing a plurality of images of a first object within a second object using near infrared light rays; capturing a plurality of images of the second object using visible light rays; extracting a plurality of silhouette points of the second object from the plurality of images on the second object; determining a set of reference points from the plurality of extracted silhouette points; detecting an angle of the set of reference points in relation to a reference line in a transverse direction of the plurality of images of the second object; aligning one or more of the plurality of images of the first object with the plurality of images of the second object based on the detected angle; detecting a plurality of projection regions of the first object when individual images of the plurality of images of the first object are projected to a projective space from different viewpoints of the individual images, the individual images being captured from a circumference of the second object; wherein the projection regions are projected in an area from a surface of projection in the projective space to a projective plane that is distant from the surface of projection by a predetermined length in a depth direction; and extracting a common portion of the plurality of detected projection regions.