目前在互联网上有海量的室外场景照片,它们不仅内容覆盖面广,对同一场景有大量不同视点和光照条件下的采样,而且获取成本很低.如何利用这些照片中包含的丰富场景信息,快速、方便地构造各种逼真的虚拟场景,是互联网迅猛发展给虚拟现实、计算机图形学和计算机视觉带来的新的研究课题之一.文中分析、总结了近年来国内外互联网图像的真实场景几何建模、自然光照建模和材质反射属性建模的最新研究进展,并对其未来的发展趋势提出了一些看法.
Photo collections are easily accessible on the Internet today, including a wide range of outdoor scenes. For each scene, the collections contain a large amount of samples with different viewpoints and illumination conditions. Exploiting the rich information in this new type of data source for the purpose of constructing realistic virtual scenes is a challenging problem in computer graphics, computer vision and virtual reality. This paper reviews recent research progresses regarding imagebased geometry modeling, illumination modeling and reflectance modeling techniques applicable to Internet photo collections, and attempts to suggest some future research directions in this domain.