Monocular vision-based 3D reconstruction of aero-engine blade damage

Aero-engine blades are one of the most important components of an aero-engine and reducing the cost of repair and maintenance of aero-engine blades is of great economic importance to the aviation industry. Borescope Technology is the most used NDT method in aero-engine overhaul, but it relies too much on the experience of maintenance personnel. We propose to combine 3D reconstruction with Borescope Technology, based on the borescope image, using monocular stereo vision principle, using SIFT algorithm for feature point detection and feature point matching by improved FLANN algorithm, using incremental SFM algorithm for sparse reconstruction of the damaged blade sequence, using PMVS method for dense reconstruction, Poisson reconstruction to enhance the reconstruction effect, and finally get The three-dimensional reconstruction model of blade damage, accurate detection of blade damage, the reconstructed three-dimensional model is the basis of aero-engine overhaul intelligence.