基于虚拟现实的大视差图像网格优化拼接算法

Grid Optimal Stitching Algorithm for Large Disparity Images Based on Virtual Reality

图像亮度、局部区域信息强度影响图像清晰度,导致图像拼接效果不佳,为提升图像拼接效果,提出一种基于虚拟现实的大视差图像网格优化拼接算法。基于虚拟现实技术建立一个虚拟现实拼接环境,在此环境下开展图像拼接处理,对待拼接图像实施灰度化处理,增强图像局部区域信息强度;基于投影算法建立低密度的网格,并根据待拼接图像的匹配点分布结果建立网格矩阵,结合全局最佳相似变换矩阵实施矩阵加权叠加,对图像重叠区域展开失真校正;使用内容感知算法对图像重叠区域展开感知识别,找出其中的积累最小像素线,实施图像融合处理,通过融合结果完成大视差图像网格优化拼接。实验结果表明,使用该方法开展图像拼接时,图像拼接时间较短,均方根误差较低,拼接效果较好。

Image brightness and local information intensity may affect the image clarity.In order to improve the image mosaics,an optimal mesh stitching algorithm for big disparity image grid based on virtual reality was proposed.Based on virtual reality,a stitching environment was constructed.In this environment,the image mosaic was carried out,and these images were grayed to enhance the information intensity of local areas.Then,low-density grids were constructed by the projection algorithm.Meanwhile,a grid matrix was constructed according to the distribution of the matching points in the images.Moreover,the matrix was weighted and superposed in combination with the global best similarity transformation matrix,and then the distortion correction was performed on the overlapping areas.The content sensing algorithm was applied to the overlapping areas of the image,thus finding out the minimum accumulated pixel line.Finally,image fusion was implemented,so that the mesh optimization splicing of big disparity images was completed.Experimental results show that the proposed method needs a shorter image mosaic time,and has a lower root mean square error,with better mosaic effect.