Research On Key Technologies Of Aero-Engine Blade Inspection Based On Robot Visual Positioning

Aero engines work in a high-temperature and high-pressure environment for a long time,and key components such as compressor blades are prone to cracks,notches and pits,which in turn cause greater safety problems.The use of industrial endoscopes to penetrate deep into the engine,or the use of eddy current testing after removing parts,not only greatly depends on the inspector’s experience level,but also easily causes misjudgment after working for a long time.The application of vision to the robot,so that the robot can perceive the information from the outside world and feed it back to the controller,its application range is very wide,such as equipment manufacturing,national defense research,civil aviation.Among them,robot vision positioning technology plays an important role,robot vision to achieve the recognition of objects,so as to carry out subsequent operations.Robot visual positioning is an important part of the key technology of aero engine blade detection of robot visual positioning,object recognition and pose estimation as the "eyes" in the system,when working,it can scan and generate a three-dimensional model of the blade surface,and then transmit the model information to the computer terminal to realize the inspection of aero engine blades.Starting from robot visual positioning,this thesis proposes a C-S segmentation algorithm under LSFP to segment scene point clouds,aiming at the problems of inaccurate positioning of blades in robot visual positioning and low efficiency in additive repair environment,and studies the hand-eye calibration and positioning effect involved.An improved registration algorithm based on key point boundary enhancement is proposed to locate the blade position in the experimental environment,and a quantitative analysis method for damage inspection of aero engine blades is proposed,which reduces the difficulty of additive blade repair in the later stage.The positioning and analysis experiments were carried out with the detection system composed of robotic arm,depth camera and 3D scanner,and the proposed algorithm was verified,and finally the damage part was analyzed and evaluated.The main contents of this thesis are as follows:(1)Introduce the research status of aero engine blade inspection based on robot vision positioning;(2)Introduce the algorithm principle of robot vision positioning and the algorithm principle of blade detection;(3)Introduce the CS segmentation algorithm flow based on LSFP in detail.The algorithm flow of improved registration algorithm based on E-ISS boundary enhancement is introduced in detail,and compared with the mainstream registration algorithm.(4)The blade inspection system under LSFP is used to carry out the calibration experiment and measurement experiment of the calibration method proposed in this thesis.According to the functional requirements of the software needed for the calibration,the system is developed based on the Qt program framework to realize the intelligent and visual positioning and inspection process.Based on the results of LSFP,a damage analysis experiment is carried out to verify the effectiveness of the proposed segmentation algorithm,improved registration algorithm and blade inspection method.