基于狭缝图像的三维虚拟漫游

3D navigation based on slit images

基于狭缝图像的四维光场模型,设计并实现了一个完整的三维虚拟场景漫游平台系统。使用四叉树结构对场景进行组织,实现了对三维场景的自适应非均匀采样,提高了采样的效率;采用临近点插补方法,解决了重采样中出现的图像空洞;通过对漫游时的内存管理、路径预测和碰撞检测处理,提高了漫游的稳定性与连贯性;对漫游速度做了测量,对绘制图像的质量做了比较。

An imagebased rendering system is developed to navigate in a virtual scene interactively. An adaptive nonuniform sampling method of a 4D light field is implemented based on slit images. The approach uses quadtree structure to reduce storage cost and accelerate the sampling process. A holefilling algorithm is designed to eliminate hole blemishes. Memory management, track prediction and collision detection are also implemented to increase the system's stability and interactivity. Experimental results on navigation speed and image quality are obtained and compared with other algorithms.