Research On Identification And Positioning Technology Of Aero-Engine Blade Grinding Area Based On Machine Vision

Aero-engine blade as the key components of the aero-engine,long-term service in high temperature,high pressure and a variety of high load conditions,very easy to appear a variety of surface defects,thus no longer meet the performance requirements of the engine.At present,China’s blade surface damage repair means mainly use manual polishing repair process,there are polishing to remove the thickness is not easy to control,repair precision is poor,low efficiency,workers labor intensity and other problems,and the process is not suitable for large volume repair.Adaptive belt grinding is a modern blade finishing technology,which can achieve highprecision repair of blade surface damage by adaptively controlling the grinding volume and planning the grinding path,and has great advantages in extending the blade service life,ensuring the blade repair quality,reducing the labor intensity of workers and improving the repair efficiency.However,the machining range of adaptive belt grinding depends on the workers’ estimation in the theoretical numerical model of the blade,which is prone to the problems of over or under regrinding area,resulting in higher machining cost or incomplete grinding repair,etc.Therefore,there is an urgent need for the technology that can obtain reasonable machining range.In this context,this paper combines machine vision,binocular vision system and surface structure light theory to determine the mapping relationship between the pixel coordinates of the area to be repaired on the blade surface in the two-dimensional image and the real three-dimensional coordinates,and then calculates the parameterized coordinate range of the defect in the theoretical numerical model of the blade according to the optimal matching algorithm to realize the identification and positioning of the blade regrinding area and provide the adaptive belt grinding parameters for the subsequent blade To provide processing range parameters for subsequent blade adaptive belt grinding.(1)To address the problem that the texture-free area of the blade profile is difficult to reconstruct reliably,this paper investigates the 3D measurement based on binocular vision system and the theory of surface structured light,and proposes a blade profile measurement method based on binocular sinusoidal structured light,which provides reliable phase features for the texture-free area of the blade by projecting a series of encoded structured light patterns onto the blade surface,and by the phase unfolding algorithm,combined with the binocular vision system The 3D point cloud of the large area of texture-free area on the blade surface can be obtained,and the reliable reconstruction of the texture-free area on the blade surface can be realized.(2)To address the problem of low imaging quality and difficult detection of defects in the resharpening area caused by blade surface reflections,this paper introduces high dynamic range imaging technology to reshape the dynamic range of the blade and alleviate the situation that the surface image features are easily lost due to blade reflections;taking blade surface scratches as an example,we propose a pre-processing process for blade surface scratch detection and a blade surface scratch classification method based on unbiased line detection and support vector machine to realize the surface scratch classification method in two-phase vision.(2)The pre-processing process and the blade surface scratch detection method based on the unbiased line detection and support vector machine are proposed to achieve the scratch detection in two-dimensional pixel scale.(3)For the positioning of the defect measurement data in the blade theoretical model,this paper firstly establishes the rigid transformation matrix between the measurement coordinate system and the design coordinate system through three sets of corresponding points to realize the basic correction of the measurement data position;then,through the accurate matching algorithm,the corresponding points of all measurement data in the blade theoretical model are obtained;finally,the parametric coordinates of the corresponding points are obtained according to the B-sample surface theory,and then the parametric coordinates of the measurement points are obtained.Finally,the parametric coordinates of the corresponding points are obtained according to the B-sample surface theory,and then the parametric coordinates of the measurement points are obtained to solve the defect positioning problem.