Master-Slave Robotic Camera Systems for Concurrent Videos in Sporting Events

点击翻译按钮获取中文摘要

Currently, live sporting events require several camera perspectives that are manually operated. The ability tocoordinate many cameras remotely and automatically will allow for more control over events are broadcasted,improvement of the reliability and responsiveness of the camera views, reduce operational costs, and ultimatelyimprove viewer experience. This project is a collaborative project between Carleton University's AdvancedBiomechatronics and Locomotion Laboratory and Ross Video, a global provider of advanced audio-visual androbotic systems to TV industry. The main objective of the project is to design and implement a master-slavecamera system in which the slave cameras each with a 3-DOF pan-tilt-zoom actuation are controlled toautomatically track the region of interest defined by the master camera (controlled by the operator) fromdifferent vantage points. There are three main aspects of this potential product that requires research: 1)Deployment and calibration of a multi-camera distributed actuation systems in large venues 2) Estimation ofdynamic desired region(s) of interest for live sport events in real-time and 3) Automate the control of motionplatforms and zoom for every camera in a network system. The project starts off by detailed evaluation of therequirements for sporting events and proposing automated camera calibration and localization methodology;developing mathematical models of the venue, cameras, and motion platforms; development of the trackingalgorithm and inverse kinematics verified via simulations; design and implementation of a room-scale setupand industrial grade infrastructure; and finally execute validation experiments. This project will lead to a newproduct for Ross Video leading to economical benefits, increased export, and high-tech job creation.