离散采集结合自适应窗截取的立体元图像生成

Elemental image generation combing discrete viewpoint pickup with adaptive window interception

为了克服离散视点采集结合窗截取的立体元图像生成方法对采集平台参数的依赖,本文提出采用基于拍摄对象的自适应窗截取算法来获得立体元图像阵列。首先,利用离散视点采集平台获得拍摄对象的离散视点图像阵列;然后,计算拍摄对象在每幅离散视点图像中的成像区域,再根据成像区域对截取窗参数的制约关系生成截取窗,并利用截取窗将离散视点图像阵列处理成子图像阵列;最后,根据立体元图像阵列与子图像阵列之间的映射关系,将子图像阵列转换成立体元图像阵列。实验结果表明:本文方法能够在采集平台参数未知的情况下,获得与传统方法相同的立体元图像阵列,大大提高了离散视点图像阵列的通用性和立体元图像阵列的生成效率。

To overcome the reliance on the pickup platform parameters of elemental image array generation method,which combines discrete viewpoint pickup with window interception,we proposed an adaptive window interception algorithm based on the object to generate an elemental image array.First,the discrete viewpoint image array is acquired using the discrete viewpoint pickup platform.Then,the imaging area of the object in every viewpoint image is calculated.Then,the interception window is generated according to the restriction on the parameters of the interception window from the imaging area,and the discrete viewpoint image array is processed into a sub-image array with the interception window.Finally,the sub-image array is transformed into the corresponding elementalimage array.Experimental results show that,without the pickup platform parameters,the proposed method can generate the same elemental image array as the traditional method can,which will boost the universality of the discrete viewpoint image array and the generation efficiency of the elemental image array.