基于计算机视觉的虚实场景合成方法研究与应用

Research and Application of Method for Synthesizing Virtual and Real Scene Using Computer Vision Technique

开发了一种虚拟场景与实时视频之间的合成技术,成功地将该技术应用于虚拟规划系统中,详细介绍了系统所采用的基于计算机视觉的标识识别和实时、自动摄像机位置、姿态跟踪算法,并给出了系统运行结果。

This paper introduces a new method to synthesize virtual and real scene using computer vision technique. The method is used in virtual plan systems. After introduce the system architecture and workflow, it mainly discusses the marker recognition method and calibration-free registration method.