【作者】 昌娅； 胡卫明； 谭铁牛；

【Author】 Chang Ya Hu Weiming Tan TieniuThe National Laboraton of Pattern Recognition, Institute of Automation, Chinese Academy of Sciences, Beijing, 100080

【机构】 中国科学院自动化研究所模式识别国家重点实验室；

【摘要】 三维模型可视化研究是视觉监控系统的一个重要组成部分。它包括建模、投影和消隐三个步骤。其中消隐又包括自消隐和物体之间的相互消隐。在实时的视觉监控系统中,对计算速度还有很高的要求。本文介绍了一种交通视觉监控系统中对三维车辆线框模型可视化的改进算法。在为车辆模型建立合理的数据结构后,算法对轮廓线和内部线做了区分,并加入了深度检验和包围盒检测,大大减少了运算量,提高了计算速度。实验结果证明了该算法的有效性。更多还原

【Abstract】 3D model visualization is an important part of visual surveillance system. It is composed of projection and clipping, including self-clipping and inter clipping among multi-models. In a realtime system calculation speed is also one of critical requirements. This paper presents an improved algorithm of 3D model visualization in the special environments for traffic visual surveillance. First we built a simple but accurate wire-frame model for the car, then after detailed analysis, we found that the lines of the model are composed of contour lines and inner lines. By making distinction between them and examination of depth and bounding boxes, our algorithm can greatly reduce the calculation in projection and clipping. These ideas are adopted in the design and implementation of a visual surveillance system. The experimental results demonstrate that the improved algorithm can greatly improve the performance of the system and meet the real-time requirement well.更多还原