Self-calibration Algorithm of Kruppa Equation Based on Planar Motion Constraints

Abstract

Camera calibration is a key technology in computer vision. Traditional methods rely on calibration objects or precise control of camera motion, which limits their application in unknown environments. Self-calibration techniques estimate camera intrinsic parameters using geometric constraints from multiple uncalibrated images. Methods based on Kruppa equations have been widely studied due to their rigorous mathematical foundation. However, these methods face challenges such as complex nonlinear solutions and sensitivity to noise in practical applications. This paper proposes an improved algorithm based on planar motion constraints, which transforms the traditional nonlinear optimization problem into solving a system of linear equations by analyzing the geometric properties of Kruppa equations under specific motion patterns, significantly reducing computational complexity. Experimental results show that this method accurately estimates camera intrinsic parameters in both simulated and real image sequences and demonstrates good robustness to noise.