Calibration strategy of optical measurement network for large-scale and shell-like objects

Abstract

It can be difficult to calibrate the three-dimensional (3D) optical measurement network (OMN) designed to inspect large-scale and shell-like objects. One of the challenges is how to in situ build up a large and precise calibration target, which can be adapted to the desired measurement volume. In this paper, a strategy for in situ calibration of the OMN is presented. First, one of the said objects is chosen to fabricate a large-scale and shell-like calibration target thereon the coded marks are pasted and their coordinates are calculated by using a technique of auto-reconstruction. This results in a highly accurate benchmark-data-set that can cover the large-scale and shell-like measurement volume. Next, all the node 3D sensors of the OMN are calibrated with the established benchmark-data-set. Thus the extrinsic parameters of all node sensors can be unified into a common coordinate system so that the structure parameters and poses of node sensors in the OMN can be determined accurately. The proposed calibration strategy is verified by a group of experiments and a case study for inspecting a large size crucible.