An effective method for extrinsic calibration of network cameras with non-overlapping field of views

摘要

This paper proposes an effective method for determining the extrinsic parameters (the locations and orientations of the cameras in a relative coordinate system), named as extrinsic calibration, for sparse network cameras, which are commonly utilized for covering much larger area. The proposed method formulates the problem as a Maximum A Posteriori (MAP) estimation, and calculates the parameters with Newton-Raphson algorithm based on the cameras' observations of an object freely moving in the cameras' field of views. One of the method's advantages is that it significantly decreases the computation time, since a step length is properly chosen and put in Newton-Raphson algorithm. The results herein demonstrate the superior performance of the proposed method over the state-of-the-art approaches based on the classical evaluation metric in simulation experiments.