Automated calibration of smartphone cameras for 3D reconstruction of mechanical pipes

Abstract

This paper outlines a new framework for the calibration of optical instruments, in particular smartphone cameras, using highly redundant circular black‐and‐white target fields. New methods were introduced for (i) matching targets between images; (ii) adjusting the systematic eccentricity error of target centres; and (iii) iteratively improving the calibration solution through a free‐network self‐calibrating bundle adjustment. The proposed method effectively matched circular targets in 270 smartphone images, taken within a calibration laboratory, with robustness to type II errors (false negatives). The proposed eccentricity adjustment, which requires only camera projective matrices from two views, behaved comparably to available closed‐form solutions, which require additional a priori object‐space target information. Finally, specifically for the case of mobile devices, the calibration parameters obtained using the framework were found to be superior compared to in situ calibration for estimating the 3D reconstructed radius of a mechanical pipe (approximately 45% improvement on average).

Towards automatic digital documentation and progress reporting of mechanical construction pipes using smartphones

“…SfM can provide 3D reconstruction up to an arbitrary scale factor only if an initial estimate of the IOPs along with the appropriate camera model are known a priori 4 . In [26], it was observed that SfM performed using initial IOPs obtained from laboratory pre-calibration, outperformed that obtained from exchangeable image file (Exif) for pipe reconstruction from smartphone cameras. Hence, in this manuscript, the smartphone cameras are pre-calibrated using the methods provided in [26,27].…”

“…In [26], it was observed that SfM performed using initial IOPs obtained from laboratory pre-calibration, outperformed that obtained from exchangeable image file (Exif) for pipe reconstruction from smartphone cameras. Hence, in this manuscript, the smartphone cameras are pre-calibrated using the methods provided in [26,27].…”

“…Current closed-form methods for correcting the eccentricity error [53,55] require detailed object space information such as target plane normal and actual target center to correct the eccentricity. In Maalek [26], a new method was proposed, which only requires the EOPs and IOPs of the images along with the ellipse geometric parameters in two views. The aforementioned discussions can be formulated using, Algorithm 2: Scale Definition in SfM, as follows:…”

Correcting the Eccentricity Error of Projected Spherical Objects in Perspective Cameras

“…Algorithm 1 requires an estimate of the interior orientation parameters (IOPs) of the camera, c, and P. This initial estimate can be obtained from pre-calibration [8,9], manufacturer's specifications, or the Exchangeable Image File (EXIF). In addition, it is possible to use Algorithm 1 to help correct camera positions in applications involving tracking spheres using cameras.…”

“…A pre-calibrated Huawei P30 mobile phone camera was used to acquire thirty-two 4K images of the spherical target. The calibration of the smartphone camera was performed using the method and IOPs presented in [9]. The position and orientation of the camera for each image view (exterior orientation parameters) were estimated automatically (independent of the spherical target's center) using COLMAP [11], a reliable open source structure-from-motion [12] software.…”