Kinematic Calibration of the KTH Head-Eye System

Mengxiang Li, Demetrios Betsis and Jean-Marc Lavest

Abstract

In this paper, we develop algorithms and techniques for kinematic calibration of a head-eye system, which consists of a pair of stereo cameras mounted a head. The aim of the calibration is to establish relationships between the 3D head (neck) system and the two camera (eye) systems, as well as the relationship between the head system and a 3D world reference system as a continuous function of the position parameters of the system. We formulate the problem and prove the necessary and sufficient conditions for it being nadaable andCVAP/ provide closed-form solutions. We also solve the problem as a nonlinear optimization problem. For comparison we also implemented some algorithms by other researchers. Testing and comparison among all the algorithms implemented are made on both simulated and real data and results are presented here.

Full paper: (PostScript 360kb)


Mengxiang Li <mxli@bion.kth.se>