Key Measurement Technologies Of Curved Surfaces Of Large Aviation Structural Component With Laser Assisted Vision Method

Large aviation structural components are the core bearing parts of large airplanes.These parts have such characteristics as large-size,weak rigidity of thin walls,and that their profile features are mostly randomly curved surfaces.Especially for large structural components made of composite material,before being assembled,the 3-dimensional(3D)profile information of the randomly curved-surfaces must be measured precisely.Then,by comparing the 3D profile information with the mathematical models of the components,the deviations of the component profiles and avert forced assembly can be analyzed.The forced assembly is caused by assembling composite structural components directly.Forced assembly can cause internal damage of materials,affect assembly quality of airplanes seriously and threaten service safety.Therefore,randomly curved surfaces measurement is of great importance for large aviation component high-quality assembly.The randomly curved surfaces of large aviation structural components have such characteristics as large profiles and easy-deformed.The 3D profiles of measured components have to be measured on the scene with a non-contact mode method.The machine vision measurement method has the advantages of non-contact,high speed and flexibility.This method has a tremendous potential application in the field measurement of large aviation curved surface part.However,the measuring range of the machine vision measurement method is limited and the measuring precision and robustness are easily disturbed by complex field environment factors such as the obtrusive light and personnel disturbance.Therefore,the conventional machine vision measurement method can not be applied directly to the field measurement of curved surfaces of large aviation structural components.To realize field measurement for randomly curved surfaces of large aviation structural components,this paper proposes a laser assisted vision measurement method.Aiming at the key technical challenges which influence the precision of measurement for randomly curved surfaces of large aviation structural components,some researches are conducted.The researches include calibration compensation for measurement system of 3D large space,high-precision extraction of non-uniform image features in scene complex environment,and precise segregation for component boundaries.The specific research contents are as follows:(1)The model of the traditional vision measurement cannot guarantee the precision in depth of field direction so that high-precision measurement cannot be achieved in three dimensional large scale space.Aiming at this problem,a binocular vision measurement method is developed based on distortion compensation for 3D images.Based on the distortion rules of visual imaging system in 3D space and the principle of binocular vision measurement,the model of binocular vision measurement based on 3D image distortion compensation is built and propose a calibration method for distortion parameters and intrinsic parameters of binocular cameras.Experiments demonstrate that the binocular vision measurement method based on 3D image distortion compensation has a better precision under condition of large depth of field than the traditional vision measurement method.(2)Since the uneven distribution of assisted laser luminance and unclear structure cause the problem that geometric features of component curved surfaces in image sequence cannot be extracted fast and precisely,a high-precision extraction method for uneven features of component curved surface in large field of view is proposed.Based on the gray model of binocular sequence laser light stripes,the factors such as feature projection angles,distances,component curved surface features,and acquisition angles of cameras can influence the rules of assisted laser luminance distribution.Then,the characters of assisted laser luminance distribution can be confirmed.Evaluation model for laser sequence images based on Gray Reliability of Structural Similarity is built to evaluate qualities of images.Based on light intensity template,region extraction method for uneven laser light stripe is proposed.Combining with the extraction results,propose fast extraction method for laser light stripe centers,based on hierarchical processing.The proposed fast and high-precision extraction method for uneven features of component curved surfaces can effectively avoid feature extraction errors.In the meantime,the darker regions of laser light stripes are still reserved completely.The extraction precision of laser light stripe centers is enhanced,compared with the classical laser light stripe extraction method of industrial measurement.(3)To solve the problem that features of component boundaries can hardly be extracted precisely because of the very disturbing noise around the boundaries,a laser boundary extraction method based fluid feature points.By means of the fluid feature points is proposed.Based on overall profile constraint,and employing arc length method to effectively extract boundaries of laser light stripes,the accuracy of extracting component boundary features is improved greatly.To conquer the limitation of vision measurement range,a research on global measurement method based on a hybrid system is conducted.Combining with the principle of global registration,global located symbols are utilized to achieve high-precision location of measurement system for field of view scanned by assisted laser,and to guarantee large-scale global measurement precision of randomly curved surfaces.The precision of global measurement can reach 0.042%.(4)By synthetically applying the key technologies of large aviation structural component measurement,a measurement system for large aviation structural component curved surfaces based on assisted laser scanning is developed.The system can realize the measurement of a large aviation structural component with the size of more than 3 meters.The system precision has been verified by international standards for 3D vision inspection and the precision satisfies the requirement of large aviation component randomly curved surfaces.The precision of this system is 0.2mm.Utilizing the vision measurement system,the key technologies mentioned are operated in this paper,and measure large aviation structural component randomly curved surfaces on scene.The results of the experiments shows that applying the large-scale and high-precision measurement key technologies,geometric features of large aviation structural.component curved surfaces can be measured well to provide effective data for analysis of structural components assembly.