Calibration Algorithms for Panorama-Based Camera Control

Abstract

We have seen that a panorama that describes the full coverage of a robotic camera is a great user interface to control the camera. The panorama provides context for users to directly point and click on the panoramic interface to command the camera to move to where they want the camera to observe. Moreover, a panorama also serves as the collaborative workspace that allows the display and the sharing of different user requests in CTRC systems. We need to guarantee the accurate correspondence between the pixel coordinates of the panorama and the camera configurations. However, the potentiometers in robotic cameras cannot provide feedback on camera configurations with sufficient accuracy. One example is the Panasonic HCM 280 camera with built-in streaming server, 22x zoom motorized optical lens, 350° pan range, and 90° tilt range. An error of 0.5° in camera tilt position can cause a 41.67% error in coverage when a Panasonic HCM 280 camera operates at its highest zoom. This presents a new calibration problem.

This chapter was presented in part in Transactions of Belarusian Engineering Academy [66].