首页>
外国专利>
METHOD AND DEVICE FOR RESTORING OMNIDIRECTIONAL CAMERA VIEW MOTION AND ARTICLE SHAPE, PROGRAM FOR RESTORING OMNIDIRECTIONAL CAMERA VIEW MOTION AND ARTICLE SHAPE, AND RECORDING MEDIUM WITH THE PROGRAM RECORDED THEREIN
METHOD AND DEVICE FOR RESTORING OMNIDIRECTIONAL CAMERA VIEW MOTION AND ARTICLE SHAPE, PROGRAM FOR RESTORING OMNIDIRECTIONAL CAMERA VIEW MOTION AND ARTICLE SHAPE, AND RECORDING MEDIUM WITH THE PROGRAM RECORDED THEREIN
PROBLEM TO BE SOLVED: To correctly restore the temporal fluctuation of the attitude and the position, i.e., the motion of the camera viewpoint to the noise specific to the time-series image, and to restore the spatial information on an external object.;SOLUTION: A characteristic point measurement unit 2 measures the temporal fluctuation of a characteristic point from the omnidirectional image of the time series of the time-series image database 1. A measurement matrix transformation unit 3 transforms the characteristic point coordinates by an omnidirectional image surface. A matrix break-down unit 5 breaks down the transformed matrix into the motion matrix of the camera viewpoint and the shape matrix to express the appearance of the external object by the factorizing method to remove the noise. A transformed matrix calculation unit 6 obtains the transformation matrix to satisfy the set condition to regulate the motion and transforms the motion matrix. The components of the camera motion are extracted from the transformed motion matrix by a viewpoint motion extraction unit 9, and the translation of the camera viewpoint and the rotational motion thereof are restored from the transformed motion matrix by a plane translation restoration unit 10 and a rotational motion restoration unit 11, respectively. A spatial information restoration unit 8 restores the rotational motion thereof. The spatial information restoration unit 8 restores the spatial information of the external object from the components extracted by an object shape extraction unit 7 by using the shape matrix transformed by using the transformation matrix.;COPYRIGHT: (C)2003,JPO
展开▼