A New Model of RGB-D Camera Calibration Based on 3D Control Field

Abstract

With extensive application of RGB-D cameras in robotics, computer vision, and many other fields, accurate calibration becomes more and more critical to the sensors. However, most existing models for calibrating depth and the relative pose between a depth camera and an RGB camera are not universally applicable to many different kinds of RGB-D cameras. In this paper, by using the collinear equation and space resection of photogrammetry, we present a new model to correct the depth and calibrate the relative pose between depth and RGB cameras based on a 3D control field. We establish a rigorous relationship model between the two cameras; then, we optimize the relative parameters of two cameras by least-squares iteration. For depth correction, based on the extrinsic parameters related to object space, the reference depths are calculated by using a collinear equation. Then, we calibrate the depth measurements with consideration of the distortion of pixels in depth images. We apply Kinect-2 to verify the calibration parameters by registering depth and color images. We test the effect of depth correction based on 3D reconstruction. Compared to the registration results from a state-of-the-art calibration model, the registration results obtained with our calibration parameters improve dramatically. Likewise, the performances of 3D reconstruction demonstrate obvious improvements after depth correction.

Single-Shot Structured Light Sensor for 3D Dense and Dynamic Reconstruction

“…Currently, 3D dynamic reconstruction techniques mainly include time of flight (TOF) [3,4], binocular stereo vision (BSV) [5,6], and structured light (SL) [7,8]. All these techniques have advantages and disadvantages, which are as follows: (1) TOF technique: The TOF technique uses active sensing to capture 3D range data as per-pixel depths [9]. A near-infrared wave is emitted by the light source from the camera, and then its reflected wave is recorded by a dedicated sensor.…”

PL-GM:RGB-D SLAM With a Novel 2D and 3D Geometric Constraint Model of Point and Line Features

“…In other words, the plane containing the camera optical center (o') and re-projected 3D line (PQ) and another plane containing the camera optical center (o) and the 3D line (PQ) should conform to the geometric coplanar condition under the unified coordinate system. The relevant error function is expressed using (12). P s and P e are 3D endpoint coordinates of a 3D line in the reference frame coordinate system.…”

“…Given (12) and the chain derivation rule, the Jacobian matrix of the error term is presented in (13) as follows:…”

“…[ 16) X s and X e are the coordinates of line endpoints in the current camera coordinate system. After acquiring the Jacobian matrices of error terms, the error Equation ( 4), ( 8), (12), and ( 15) are utilized to establish our new optimization model of points and lines. Based on each iteration of residuals, the weight of each error term related to a line is defined using the robust kernel function such as the Cauchy function.…”

“…RGB-D camera is a new type of vision sensor with the ability to provide RGB and depth images simultaneously. Kinect-1, the first RGB-D camera, was released by Microsoft in 2009 [12] and was applied in kinds of fields widely. The advantages of the RGB-D camera make it favorable to become another feasible alternative to the monocular and stereo cameras in the visual SLAM fields [13].…”