Efficient Camera Pose Adjustment to a Mirror Array for Structured Light Field Video Acquisition

We previously implemented an inexpensive imaging system that combines a single real camera with a mirror array located along a paraboloid. It allows us to robustly acquire dynamic light fields composed of multi-view videos by providing a virtual camera array, where its viewpoints exist in the mirrors. Actually, as moving the real camera to the focus of the paraboloid, virtual viewpoints in the mirrors get equally-spaced to achieve multi-view imaging with structured disparity. In this paper, we discuss an efficient method for adjusting the pose of a single camera to acquire high quality dynamic light fields as multi-view videos. Specifically, we introduce some indicator values determined by detected corners of the mirror array on acquired images while adjusting the camera. By using these values for camera adjustment, we easily know how to move its position and virtually correct its angle through homography transform. Experimental results of simulations demonstrate that our proposed method sufficiently achieves structured light field video acquisition with equally-spaced virtual viewpoints, where we do not need camera rotation requiring complex devices and only the camera position is controlled by a simple 3D system like XYZ stages.

I. Introduction

Advanced 3D visual media often require dynamic light fields or multi-view videos for ambient communication and immersive user experience [1]–[3]. In order to enhance such media without limiting them to computer-generated synthetic ones, we need to develop technologies for acquiring dynamic light fields composed of real multi-view videos inexpensively and robustly. Actually, dynamic light fields have been obtained as multi-view videos captured by camera arrays [4]–[9]. Massive camera arrays allow us to acquire high-quality real multi-view videos of sufficient image resolution by a large number of viewpoints. However, due to many device components including video synchronization and aggregation for a camera array, it not only costs too much but also severely suffers from low fault tolerance. In contrast, we can enjoy an inexpensive light field camera combining a micro-lens array with a single image sensor to acquire multi-view still images [10]–[15]. Unfortunately, although it captures very dense light fields, its optical device causes insufficient exposure, that leads to awful noise on videos. Thus, light field cameras cannot easily obtain high-quality multi-view videos as robust dynamic light field acquisition.