Study On Background Removal And Improved Zhang Zhengyou Calibration Method For Optical3d Measurement System Of Titanium Blades

Recently the rapid development of domestic and international aviation industry makesan important contribution to the aero-engine repair technology’s influence. However, in fact,as one of the important technologies of aero-engine repair, the development of bladeswelding repair technology in China is just beginning and facing technical blockade ofwestern countries. Therefore, in order to cut down this unfavorable situation’s influence,developing Chinese aero-engine defect blades welding repair technique is significant.Before the work of welding repair, the three-dimensional precision measurements of defectblades will determine the final welding quality.The high precision and fast three-dimensional measurement for aero-engine blades in theautomatic welding repair system is this paper’s background. With the dual-frequency gratingmeasurement technique as the theoretical basis, building system measurement platform andcarefully analyzing the principle of background removal problems and system calibrationstrategy, finally proposing a two-step method for removing interference in dual-frequencyphase-expansion and a12pixels symmetric template checkerboard interior points detectionalgorithm for camera calibration.The solution phase method by dual-frequency projection grating3D measurementprinciple is widely used because of its less susceptible to noise interference. But the studyfound that when the measurement object is titanium blades, the blade’s continuous phasediagram will be indistinguishable from the background image, which is mainly becauseblade titanium body’s color is too similar to the reference plane’s color. For this problem,proposing a two-step method to remove the background: firstly, in order to increase thecolor contrast between the reference plane and titanium blade so that can effectivelydistinguish them, the reference plane is preset to be a black background; then combinedwith the Imcrop image cutting function from MATLAB software to automatically interceptthe blade deformed grating images, eventually getting the full blade continuous phasediagram. Experimental achievements demonstrate that the proposed method can separateblade’s phase map from background image effectively, while increasing the phase wrappedaccuracy.When doing camera calibration work, in order to solve calibration toolbox manualextraction checkerboard corners’ low accuracy and efficiency problem, proposing a12pixels symmetric template checkerboard interior points detection algorithm based on classicSUSAN(smallest univalue segment assimilating nucleus) algorithm for camera calibration.Experimental achievements demonstrate that the new algorithm can extract checkerboardinterior angle points accurately, and its average detection time is1.244577s for9order checkerboard corner, while decreasing corner extraction reprojection error to [0.3,0.3] pixelsby Zhang’s camera calibration method, so that the checkerboard interior points can be detectedquickly and efficiently.