Automated Borescope Defect Detection Using RPCA
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current automated inspection techniques for borescope images are prone to errors due to repetitive tasks and human fatigue, and manually set detection thresholds are error-prone, often missing critical defects in aircraft engine blades.
Innovation Solution
An automated defect detection method using Robust Principal Component Analysis (RPCA) to decompose images into low rank and sparse component images, followed by additional processing such as statistical techniques and filtering to confirm defects, minimizing human intervention and manual threshold adjustments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If human inspectors manually examine borescope images for defects, then defect detection can be performed with human judgment, but inspector fatigue and repetitive tasks lead to missed defects and reduced reliability
Solution Approach 1:
The system performs self-service by automatically detecting defects through RPCA-based image decomposition without requiring human inspectors to manually examine each image. The automated algorithm independently processes borescope images, identifies defects through mathematical decomposition, and generates inspection results, thereby eliminating human fatigue and inconsistency while maintaining high reliability.
Solution Approach 2:
The patent replaces the mechanical human inspection process with an automated computational system. Instead of human eyes and brains examining images, the system uses Robust Principal Component Analysis algorithms to automatically decompose images and detect defects, substituting mechanical human labor with an automated electronic processing system that maintains consistent reliability across all inspections.
2Extent of automation
If manually-set detection thresholds are used in automated inspection systems, then automation is achieved, but the system becomes error-prone and cannot detect defects outside predefined classes
Solution Approach 1:
The patent changes the fundamental parameter of defect detection from threshold-based classification to decomposition-based identification. Instead of using fixed detection thresholds that require manual setting and cannot adapt to new defect types, the system uses RPCA to decompose images into low-rank and sparse components, where defects naturally emerge as sparse anomalies. This parameter change enables the system to detect any defect type without predefined thresholds, improving both automation and reliability.
3Measurement precision
If focus is placed on low-level feature extraction and matching to predefined defect classes, then common defects can be identified, but the system cannot detect defects outside those pre-defined classes
Solution Approach 1:
The patent applies universality by creating a single RPCA-based detection system that can identify all types of defects regardless of their specific characteristics. The low-rank sparse decomposition approach is universally applicable to any defect type, whether cracks, erosion, nicks, or previously unknown defect forms. This universal method eliminates the need for separate detection algorithms for different defect classes, thereby achieving both precision and broad adaptability.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
[0035] A computer program product and method for performing automated defect detection of blades within an engine is disclosed. The method may include providing a storage medium for storing data and programs used in processing video images, providing a processing unit for processing images, receiving from a borescope an initial set of images of a plurality of members inside of a device, and using the processing unit to apply Robust Principal Component Analysis to decompose the initial set of images into a first series of low rank component images and a second series of sparse component linages, wherein there are at least two images in the initial series.