WIEPnP-based 3D expansion measurement of water walls of utility boilers

Abstract

To address the expansion failure risk of water walls in utility boilers due to flexible operating conditions, a 3D expansion measurement method based on monocular vision is proposed. This method first utilizes PnP algorithm to obtain the camera pose, and then gets the 3D expansion through matrix transformation. To reduce the impact of depth variation and image noise on the 3D expansion measurement, an innovative algorithm is developed from EPnP, as weighted iterative EPnP (WIEPnP). In WIEPnP, weight coefficients are assigned to the markers in the optimization objective function and iterative calculations are employed to minimize the impacts on the measurement results from depth variation and image noise. Experiments on a prototype measurement system were conducted in laboratory. The results convinced high accuracy of the WIEPnP-based expansion measurement system, in X, Y, and Z directions the displacement measurement errors are all less than 0.8 mm. Additionally, time for a single measurement is shorter than 2 seconds. It shows that the proposed WIEPnP algorithm meets all the requirements for real-time expansion measurement of water walls of utility boiler.