Automatic inspection system of aeroengine cooling hole based on gas excitation and infrared images

Abstract Film cooling holes are critical for reducing surface temperature and prolonging the service life of turbine blades in aeroengines. These holes are prone to defects such as shrinkage cavities and blockages after coating or long-term use. To address this, we propose an automated inspection system based on infrared image sequences generated by gas thermal excitation, integrating a robotic arm, an infrared camera, and deep learning techniques. A key innovation of our system is a dedicated image preprocessing pipeline designed to enhance the detectability of film cooling holes. This includes temporal differentiation to highlight dynamic thermal behavior, and local patch-based normalization to amplify grayscale contrast between the holes and their surroundings, thereby improving detection sensitivity and robustness. The system workflow includes: (1) analyzing the response of different blade types to various excitation signals; (2) collecting infrared sequences with a robotic arm to build a comprehensive dataset; and (3) applying deep learning algorithms, combined with the preprocessing techniques, for accurate localization and diameter estimation. Comparative experiments confirm the superiority of our approach over traditional inspection methods.