Multiple Static Cameras Research Articles

LEDs based video camera pose estimation

Abstract For 3D object localization and tracking with multiple cameras the camera poses have to be known within a high precision. The paper evaluates camera pose estimation via a fundamental matrix and via the known object in environment of multiple static cameras. A special feature point extraction technique based on LED (Light Emitting Diodes) point detection and matching has been developed for this purpose. LED point detection has been solved searching local maximums in images and LED point matching has been solved involving patterned time functions for each light source. Emitting LEDs have been used as sources of known reference points instead of typically used feature point extractors like ORB, SIFT, SURF etc. In such a way the robustness of pose estimation has been obtained. Camera pose estimation is significant for object localization using the networks with multiple cameras which are going to an play increasingly important role in modern Smart Cities environments.

A visualization framework for team sports captured using multiple static cameras

We present a novel approach for robust localization of multiple people observed using a set of static cameras. We use this location information to generate a visualization of the virtual offside line in soccer games. To compute the position of the offside line, we need to localize players’ positions, and identify their team roles. We solve the problem of fusing corresponding players’ positional information by finding minimum weight K-length cycles in a complete K-partite graph. Each partite of the graph corresponds to one of the K cameras, whereas each node of a partite encodes the position and appearance of a player observed from a particular camera. To find the minimum weight cycles in this graph, we use a dynamic programming based approach that varies over a continuum from maximally to minimally greedy in terms of the number of graph-paths explored at each iteration. We present proofs for the efficiency and performance bounds of our algorithms. Finally, we demonstrate the robustness of our framework by testing it on 82,000 frames of soccer footage captured over eight different illumination conditions, play types, and team attire. Our framework runs in near-real time, and processes video from 3 full HD cameras in about 0.4s for each set of corresponding 3 frames.