A flexible calibration method of multiple Kinects for 3D human reconstruction

Abstract

In this paper, we present a simple yet effective calibration method for multiple Kinects, i.e. a method that finds the rel ative position of RGB-depth cameras, as opposed to conven tional methods that find the relative position of RGB cameras. We first find the mapping function between the RGB camera and the depth camera mounted on one Kinect. With such a mapping function, we propose a scheme that is able to esti mate the 3D coordinates of the extracted corners from a stan dard calibration chessboard. To this end, we are able to build the 3D correspondences between two Kinects directly. This simplifies the calibration to a simple Least-Square Minimiza tion problem with very stable solution. Furthermore, by using two mirrored chessboard images on a thin board, we are able to calibrate two Kinects facing each other, something that is intractable using traditional calibration methods. We demon strate our proposed method with real data and show very accu rate calibration results, namely less than 7mm reconstruction error for objects at a distance of I.Sm, using around 7 frames for calibration.

Multiview 3D sensing and analysis for high quality point cloud reconstruction

“…The ICP calibration applied in [2] and [5] shows good result with overlapping areas of point clouds. The method presented in this paper is based in the classic planar pattern detection in 3D space described in [4].…”

“…For each device, an extrinsic matrix has to be calculated. Following the approach presented in [4], we calibrated the relative pose between Kinect v2 sensors using the IR camera of each Kinect. In Figure 3, the captured IR images from Kinect v2 sensors with a standard chessboard planar object is depicted.…”

“…Related approaches have been extensively studied and reported in the literature. For instance, in [4] and [5] a multi-view sensing solution with Kinect v2 devices using separate computers connected to a manging server is presented. However, after the initial point cloud capturing complex post-processing stages are required to achieve high quality representations.…”

Bi-objective Framework for Sensor Fusion in RGB-D Multi-View Systems: Applications in Calibration

“…The scanning results are promising, remark that, however, the obtained 3D reconstructions are rigid, even if the objects in the environment were constantly moving and deforming. This issue can be solved by using multiple fixed RGB-D cameras covering the entire scene [18,20,11,28]. In this case, the relative poses of all RGB-D cameras are required for aligning the partial 3D reconstructions.…”

Multilevel active registration for kinect human body scans: from low quality to high quality

“…Current publicly available high-quality human body datasets, such as SCAPE [3], FAUST [5], TOSCA [8] are built either from costly laser scanners or need other assistance (e.g makers, texture or professional tools). With the appearance of low-cost scanners such as Kinect, it is now possible for an object, a room or even a person to be quickly scanned, modeled and tracked [12,14,27,28,30,31,44]. Nowadays human body meshes could be captured for different identities in different poses in a few minutes.…”