发明名称 Stereoscopic image generating device and stereoscopic image generating method
摘要 A stereoscopic image generating device includes: a correction parameter calculating unit that calculates correction parameters based on a plurality of pairs of feature points corresponding to the same points on the object, from the first image and a second image photographing the object; a correction error calculating unit that, for each pair of feature points, corrects the position of the feature point on at least one image, using the correction parameters, and calculates the amount of correction error; a maldistribution degree calculating unit that finds the degree of maldistribution of feature points; a threshold value determining unit that determines a threshold value such that the threshold value is smaller when the degree of maldistribution increases; and a correction unit that, when the amount of correction error is equal to or lower than the threshold value, corrects the position of the object in the images using the correction parameters.
申请公布号 US9235897(B2) 申请公布日期 2016.01.12
申请号 US201213606445 申请日期 2012.09.07
申请人 FUJITSU LIMITED 发明人 Kakuko Norihiro;Sato Teruyuki
分类号 G06K9/46;G06T7/00;H04N13/00;G06K9/00 主分类号 G06K9/46
代理机构 Fujitsu Patent Center 代理人 Fujitsu Patent Center
主权项 1. A stereoscopic image generating device comprising: a feature point extracting unit that extracts, from a first image that is generated by photographing an object by a first camera, and a second image that is generated by photographing the object by a second camera, a plurality of pairs of feature points corresponding to same points on the object; a correction parameter calculating unit that calculates at least one correction parameter to correct deviation between a position of the object on the first image and the position of the object on the second image, based on the plurality of pairs of feature points; a correction error calculating unit that, for each of the plurality of pairs of feature points, corrects a position of the feature point of at least one of the first image and the second image using the correction parameter, finds a distance between feature points corresponding to same point on the object after the correction, and calculates an amount of correction error which represents a statistic of the distance; a maldistribution degree calculating unit that finds a degree of maldistribution of feature points by dividing the first image or the second image into m in the horizontal direction and n in the vertical direction, thereby setting up m×n blocks, m and n being an integer and at least one of m and n being greater than 1, and calculating, for each block, the number of feature points included in the block, and using the number of blocks including no feature points or including a smaller number of feature points than a predetermined value, as the degree of maldistribution; a threshold value determining unit that sets the amount of correction error as of when a probability that an adequate stereoscopic image is generated by performing correction using the correction parameter, is a predetermined value, as the threshold value, based on a relational equation to represent a relationship between the probability and the amount of correction error, determined per degree of maldistribution; a decision unit that discards the correction parameter when the amount of correction error is greater than the threshold value; and a correction unit that, when the amount of correction error is equal to or lower than the threshold value, generates a stereoscopic image by correcting at least one of the position of the object on the first image and the position of the object on the second image using the correction parameter.
地址 Kawasaki JP