基于VR技术的大视角成像图像拼接仿真  

Simulation of Large View Imaging Image Mosaic Based on VR Technology

作　　者：马璿 陈烽[1] MA Xuan;CHEN Feng(Department of Information Engineering,Tibet University for Nationalities,Xianyang Shaanxi 712000,China)

机构地区：[1]西藏民族大学信息工程学院,陕西咸阳712000

出　　处：《计算机仿真》2022年第7期213-217,共5页Computer Simulation

基　　金：西藏民族大学“青年学人”资助项目(21MDX02);西藏自治区自然科学基金项目(XZ202001ZR0060G)。

摘　　要：由于成像图像拼接过程中拼接重叠区域特征点歧义性高、匹配度较低,导致成像图像拼接难度较大。当前方法存在畸变、视觉观感差的问题。提出新的基于VR技术的大视角成像图像拼接方法。通过函数测算、变换选择特征点和待匹配点明确待拼接图像内特征点集合,利用图像间重叠区域位移值,计算离散灰度与相关函数估计值,得到图像间重叠特征点提取结果。通过相似度和匹配度将各个匹配点集匹配,并建立单应性矩阵。将首个图像作为基准,剩余图像则在首图像基础上延伸串联,同时加权处理水平与垂直方向的像素,实现大视角成像图像拼接。仿真结果表明,所提方法能够提高匹配特征点精度和匹配度,拼接后图像无歧义且过渡平滑,视觉观感佳。Due to high ambiguity and low matching degree between feature points in the overlapping region during the image mosaic,it is difficult to stitch images.Therefore,a method of image mosaic for large viewing angle imaging based on VR technology was presented.Through the function calculation,transformation and selection of feature points and the points to be matched,we defined the set of feature points in the image to be spliced.Then we used the displacement value of overlapping regions between images to calculate the estimation values of discrete gray levels and correlation functions,and thus to get the results of extracting overlapping feature points.Moreover,we matched all sets of matching points by similarity and matching degree,and built the homography matrix.Furthermore,we took the first image as the benchmark,and made the remaining images extended in series.Meanwhile,we weighted the pixels in horizontal and vertical directions,thus realizing the image mosaic for large viewing angle imaging.Simulation results prove that the proposed method can improve the accuracy and matching degree of matching feature points.In addition,the images are unambiguous after splicing,and the image transition is smooth,so the visual perception is good.

关 键 词：图像畸变 特征点匹配校准 图像拼接 单应性矩阵 离散灰度 

分 类 号：TN912[电子电信—通信与信息系统]