Abstract

Fast development in image processing and classification techniques brings many new solutions to the challenges in engineering. In this thesis, image processing and classification techniques are introduced as a powerful inspection tool into non-destructive testing on aircraft parts and honey bee disease detection to improve inspection speed and accuracy of human inspectors.
Safety and reliability are the most important issues in aerospace industry, especially in high temperature and pressure turbine engine parts. Fluorescent Penetrant Inspection (FPI) is widely used in Non-Destructive Testing (NDT) on aircraft parts as an easy and powerful method. To improve efficiency and robustness in human inspection in FPI, an Advanced Automatic Inspection System (AAIS) is developed by using image processing and classification techniques in this thesis. The system can automatically detect, measure and classify the discontinuities from turbine blade FPI images.
As the world’s twelfth-largest honey producer, Canada’s honey bee suffers from big losses in the past few years. The detection of disease and disorder of bee colony in an early stage is critical to prevent more loss for the bee industry. To greatly improve the speed of the inspection while retain accuracy, an automatic health monitoring system is developed to inspect honey bee colony and detect disease and disorder. The system can monitor colony development and measure the proportion of unhealthy cells. And thus it can improve the efficiency and accuracy of bee colony inspection significantly.