Deep Learning based Framework for Automatic Damage Detection in Aircraft Engine Borescope Inspection

Borescope inspection has limited accessibility, creating the need to use different borescope tips, which leads to high variation of images.

“Furthermore, there is limited accessibility, which creates the need to use different borescope tips, which in turn leads to a high variation of images [9].”

There is a lack of recent applications of classical image processing techniques to blade inspection.

“There is a lack of recent applications of these techniques to blade inspection. In fact, the field is somewhat weak and has been dominated by the neural networks and deep learning approaches instead [9,19,47,85].”

CNN approaches for crack and erosion detection typically achieved high accuracy of detection.

“All of these approaches use some form of feature extraction or segmentation techniques to normalise the input into a trained CNN and typically achieved a high accuracy of detection.”

All these CNN applications required large quantities of training data.

“Additionally, all these applications required large quantities of training data.”

The training examples used in CNN approaches were of advanced damage that is clearly visible to the human eye.

“However, the training examples were of advanced damage that is clearly visible to the human eye.”

Engine parts such as shafts, fan blades, compressor blades and turbine blades are safety-critical to flight operations.

“Engine parts such as shafts [31,32], fan blades [20,33], compressor blades [9,34,35] and turbine blades [9,26,36-43] are also of particular interest as all are safety-critical to flight operations.”

Most research focuses on defect detection on turbine blades rather than compressor blades, mainly crack detection.

“Furthermore, most research focuses on defect detection on turbine blades rather than compressor blades. This encompasses mainly crack detection [9,33,37,38,42], as this is the most critical type of defect and the main source of failure.”