Accurate Method for Real-time Defect Detection

Back to all technologies
Download as PDF
2019-JAHA-68509
Equipment inspections are necessary for quality control and maintenance to keep equipment running safely. However, current inspection methods can involve time-consuming and costly methods where a human inspector must accurately and quickly identify defects on a surface. Researchers at Purdue University have developed a defect detection approach to inspect the internal components faster and more accurately than current industrial and other state-of-the art practices. This approach provides precise crack contours that are useful in assessing the severity of cracks while achieving 98.5% hit rates against 0.1 false positives per frame. Furthermore, this approach is 110 to 150 times faster when compared to other state-of-the art approaches. This new approach is a significant step toward real-time video analysis for autonomous defect detection.

Advantages:
- Accurate
- Fast
Potential Applications:
- Video Analysis for Defect Detection
Apr 9, 2019
Provisional-Patent
United States
(None)
(None)
Purdue Office of Technology Commercialization
1801 Newman Road
West Lafayette, IN 47906

Phone: (765) 588-3475
Fax: (765) 463-3486
Email: otcip@prf.org