Share Email Print

Proceedings Paper

Aircraft engine blade cooling holes detection and classification from infrared images
Author(s): Robert D. Rosemau; Sal Nawaz; Aiqun Niu; William G. Wee
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

In order to inspect sets of cooling holes being properly drilled on an aircraft engine blade, a robotic controlled IR imaging system was designed. The system physically having a robot held the drilled blade at different viewing positions in front of an IR camera where a sequence of images were obtained for analysis and evaluation where a hot-air cool- air heating and cooling cycle was being administered at the base of the blade. An initial teach model was needed to enable the robot to remember the different viewing position of the blade in front of the camera so as to establish a position and orientation reference and sequence of cooling hole inspection. Two data processing algorithms had been achieved in this paper: 1) A position adjustment during the regular operation after the tach-mode so as to made sure that he cooling holes to be inspected matched the reference established by the teaching phase; 2) An image processing method to extract four meaningful features and to drive a pattern recognition system for the determination of cooling hole operating. Blade classification was based on the number of bad cooling holes as being good if all but one hole is determined to be good, unless the bad hole is positioned in one of four critical locations, in which case the blade was classified as bad. A hole is classified as being bad if the hole is blocked in some manner, or if the hole's diameter does not fall within a specified range. Two types of data which were considered to be significant in characterizing the hole were the temperature intensity of the hole as it was heated and cooled, and the change in the intensity as a function of time during the same period. Image data features for the holes were extracted from the images and applied to several classifiers to determine an optimal classification method. Different data sets and/or combinations of features for both training and testing sets were formed and tested. Over 90 percent performances were achieved under different evaluation methods.

Paper Details

Date Published: 28 January 1999
PDF: 9 pages
Proc. SPIE 3586, Nondestructive Evaluation of Aging Aircraft, Airports, and Aerospace Hardware III, (28 January 1999); doi: 10.1117/12.339875
Show Author Affiliations
Robert D. Rosemau, Univ. of Cincinnati (United States)
Sal Nawaz, Univ. of Cincinnati (United States)
Aiqun Niu, Univ. of Cincinnati (United States)
William G. Wee, Univ. of Cincinnati (United States)

Published in SPIE Proceedings Vol. 3586:
Nondestructive Evaluation of Aging Aircraft, Airports, and Aerospace Hardware III
Ajit K. Mal, Editor(s)

© SPIE. Terms of Use
Back to Top