Strategies for generating panoramic video images without information about scene correspondences for multispectral distributed aperture systems

Abstract

We derive analytical expressions for calculating the number of elementary computational operations required to generate several personal regions of interest in a panoramic computer-vision distributed-aperture system using two alternative strategies: strategy 1 involves acquisition of a complete panoramic frame, followed by the selection of personal regions of interest, while with strategy 2 the region of interest is directly formed for each user. The parameters of analytical expressions include the number of cameras in the distributed system, the number of users, and the resolution of panorama and user frames. The formulas obtained for the given parameters make it possible to determine a strategy that would be optimal in terms of a criterion of the minimum number of elementary computational operations for generating multiple personal regions of interest. The region of interest is generated using only a priori information about the internal and external camera parameters, obtained as a result of their photogrammetric calibration with a universal test object, and does not take into account information about scene correspondences at the boundaries of intersecting fields of view.

Simulation on Basketball Recognition Based on Mobile Network Communication and Panoramic Vision

“…rough the experimental design in the previous section, two aspects are mainly involved. On the one hand, it is the choice of the preprocessing method for converting the 2D coordinates of the video into 3D coordinates, and on the other hand, it is the use of the action recognition algorithm [24]. By inputting a basketball motion video, the 2D skeleton information is first extracted from the classified video, and then, the 3D skeleton coordinate information is matched in real time through videoPose3D.…”