Accurate Plane Estimation Based on the Error Model of Time-of-Flight Camera

Abstract

In this paper, we propose an accurate plane estimation method for 3D point cloud data acquired by ToF cameras. The proposed method is based on an error model of the ToF camera derived from its acquisition geometry and noise characteristic. We utilize the distance between point and plane along the ray direction; the ray-directional distance. We give a brief approximation of a noise model for the ToF camera and utilize it as an adaptive weighting on the ray-directional distances. We formulate the plane estimation problem as the minimization of weighted ray-directional distance. Experimental results demonstrate that the proposed method can outperform conventional plane estimation methods.