A high-precision global unified method based on loop constraints for measurement with 3D scanner

Abstract In the field of 3D object vision measurement, accurately determining the transformation relationship between local and global coordinate systems is crucial. However, existing methods reduce the accuracy of global unification by ignoring global errors among coordinate system transformations. This paper solved this problem by establishing a global measurement system based on optimized coded mark points. Subsequently, a global loop constraint was constructed to calculate the pose of the scanner. Finally, the 3D point cloud of the measured object was unified globally by each pose. The effectiveness of the measuring system and method was evaluated through comprehensive experimentation. The reprojection root-mean-square error (RMSE) with global unified accuracy is less than 0.1 mm, and the RMSE of the measurement for standard ball diameter was 0.074 mm. The experimental results demonstrate that the method exhibits robust unification capabilities in the reconstruction of 3D objects.