摘要
为了获取整个球面全景图像,提出了全方位视觉传感器(omni-directional visionsensor,ODVS)的二次折反射的成像原理并设计了折反射镜面。利用4阶Runge-Kutta算法求得了镜面曲面的数值解,通过组合镜头的设计实现ODVS与广角镜头的集成以消除ODVS的固有死角,然后,将两台无死角的ODVS装置以背靠背的方式进行连接,并将所获取的全景视频图像进行展开及无缝拼接得到360°·360°视频图像。实验结果表明,所设计的ODVS装置可以获取全球面的视场范围,在视频监控领域有重要的应用价值。
In order to obtain a panoramic image with a full sphere view, this paper presents the two-catadioptric imaging principle of omni-directional vision sensor (ODVS), by which catadioptric mirrors can be made. The reflectors curvilinear numeric solutions of the catadioptric mirrors are obtained by means of the fourth order Runge-Kutta algorithm, the problem of dead angle existing in current ODVSs is solved through designing the combination lens, in which ODVSs and wide-angle lens are put together. In addition, two ODVS devices without dead angle are fixed in a back-to-back way, and then, the captured images by the two ODVS devices are unwrapped and seamlessly composed so that an image with 360° by 360° is obtained. The experimental results show that the proposed ODVS devices are capable of capturing panoramic images with a full sphere view and can be used in intelligent video surveillance fields.
出处
《高技术通讯》
EI
CAS
CSCD
北大核心
2009年第7期724-729,共6页
Chinese High Technology Letters
基金
国家自然科学基金(60673177)
浙江省科技厅重大科技(2006C11202)资助项目。