摘要
在分析了以往所采用的改进图像过分割方法的基础上,提出了一种新的虚拟场景构建算法。利用单场景建模,把图像分割成超像素,接着为超像素增加类别标志从而形成集群,然后由集群确定几何类别,最后剪切和折叠不同标志的图像,输出三维场景。该算法利用VMRL对一张图片进行建模,在虚拟漫游系统中无需用户直接参与,可以高效地构建虚拟现实场景。
Based on the improved image over-segmentation, a new virtual scene construction algorithm is proposed. Firstly, the image is segmented into several superpixels, and then the superpixels are labeled into constellations, which are specified with geometric characters. At last, the preprocessed image is cut and folded and changed into three dimensional scene. The algorithm uses one image to model, needing no user interference, and can effectively construct the virtual reality scene.
出处
《计算机工程与设计》
CSCD
北大核心
2009年第17期4044-4046,4107,共4页
Computer Engineering and Design
关键词
图像过分割
单场景建模
超像素
场景构建
虚拟现实场景
image over-segmentation
single view modeling
superpixel
scene construction
virtual reality scene
