Combining Kinect and PnP for camera pose estimation

This paper presents a novel method to conduct camera pose estimation though combining Kinect and Perspective-n-points algorithms. Most existing camera pose estimation methods suffer from the errors caused by inevitable outliers between 2D-3D correspondences. To this end, we propose to use a random down sampling process to deal with outliers in this paper. The proposed method is divided into two main steps, which are 2D-3D correspondences generation and pose estimation. The method has been tested in a real project, and the experiment has shown encouraging results compared to the ground truth.

I. Introduction

Camera pose estimation is a widely used technique for multi-camera related applications. The main objective of the estimation is to recover the Rotation and Translation of one camera in a certain coordinate system from its image. There are three kinds of implementation for this estimation process, which are the ones based on direct linear transformation (DLT), perspective-n-points (PnP), and a priori information estimator. Due to its higher robustness, PnP has wider users than others. However, to achieve accurate estimation results, PnP requires a set of high accurate 2D-3D correspondences between 2D points in camera image and 3D points in the space, which would be obtained with inevitable outliers that could leads to potential errors.