Accurate radius measurement of multi-bend tubes based on stereo vision

To realize accurate and fast radius measurement for the multi-bend tubes in the 3D environment, a stereo vision based method was approached in this paper. Firstly, the mapping from tube's center line to edge point in the image was modeled and discussed. Then, the photo-consistency was used to reconstruct 3D center line of the bend tube. Furthermore, a segmented parametric model of tubes with curved parts and straight parts was developed, and the problem was converted to a constrained optimization problem. The experiments show the measurement accuracy of this approach can reach to ±0.5 mm with 3σ reliability. The results of measurements showed good performance in terms of precision and reliability, even when lacking corresponding points or with incomplete edges. Besides, the approach can measure the bend radius automatically without artificial operations.