中文摘要：

在对全景实时处理要求下,全方位视觉传感器（ODVS）是一种快速、可靠的视觉信息采集途径,为了克服目前ODVS技术还存在着死角的问题,探讨了一种具有平均角分辨率的ODVS的二次折反射的成像原理,根据该成像原理设计了这种新型的ODVS的结构和折反射镜面,利用4阶Runge-Kutta算法求得了镜面曲面的数字解,然后根据组合镜头的设计方法将ODVS与广角镜头集成在一起来获取原ODVS技术所存在的死角视频图像。实验结果表明,本研究所设计的ODVS能实现真正意义上的全方位监控,在全方位视频智能监控、动态图像理解、计算机视觉等领域具有重要的应用价值。

英文摘要：

In real-time panoramic image processing, Omni-Directional Vision Sensor （ODVS） is a fast and reliable means of visual information collection, but current ODVS has a dead angle. According to the average angular resolution of two catadioptric ODVS imaging theory, a new design of ODVS structure and catadioptrie mirror is presented in this paper to overcome the defect. The reflector curvilinear digital solution of this catadioptric mirror is obtained using the fourth order Runge-Kutta algorithm, and the dead angle part of the panoramic image that normal ODVS cant acquire is acquired by a combined-lens of a wide-angle lens and a normal ODVS. Experimental results show that the ODVS presented in this paper can achieve true omni directional surveillance, and has a good application value in the fields such as intelligent omni directional video surveillance, dynamic image understanding and computer vision.