Automatic video mosaicing speeds surveillance time

Researchers Marcos Nieto and Luis Salgado in the Image Processing Group at the Universidad Politécnica de Madrid (Madrid, Spain) have developed a surveillance technique using moving cameras and projective geometry to enable generation of real-time, user-controlled superimages of scenes. Video surveillance typically relies on various methods of 3-D scene reconstruction, such as people counting and object detection and tracking. Video mosaicing is a related technology that enables an expanded view of a scene by pasting together frames from the video stream (superimages) of a moving single camera in real time. The researchers have tested their vanishing point estimation algorithm against the most popular and recent methods and have shown that their approach is both highly accurate and significantly reduces the cost of computation. The result is an online application that works in real time (15 frames/s) and achieves calibration on structured scenarios for moving PTZ cameras and video mosaicing. For more information, view the article on www.spie.org. Posted by Conard Holton, [email protected], Vision Systems Design