Research on the key techniques of 3D defect detection on aeroengine blade surface utilizing robotic vision

Aeroengine blade is a key component of aeroturbine, and its surface state always changes due to its extreme work environment such as high temperature and high pressure. Defect detection of the aeroengine blade is essential for turbine safety. Existing detection methods have the disadvantages of 2D detection, long time and low intelligence. In this paper, robotic vision technique is used for the intelligent 3D defect detection of aeroengine blade. An accurate structured-light vision 3D measurement sensor is developed. An accurate sphere-based calibration approach with enhanced adaptability is developed for 3D vision sensors. Then, the 3D vision measurement sensor is fixed on the robot, and the aeroengine blade defect can be detected as follows: deep learning technology is used to the captured picture by the vision sensor to recognize defects intelligently; cloud points of the the recognized defect can be achieved by the 3D vision measurement sensor, which can provide data support for blade life assessment.