MULTI-FREQUENCY IMAGE PROCESSING FOR INSPECTING PARTS HAVING COMPLEX GEOMETRIC SHAPES
Patent №
US 7,817,845
Granted
2010-10-19
Filed 2006
Owner
GENERAL ELECTRIC COMPANY
Lab
—
AI components
2
vision · planning
Assignment
Recorded
Dataset
AIPD
2023_r1 edition
Application
11617859
A method for detecting small cracks and other anomalies on parts having complex geometries is disclosed. The method includes eddy current inspection incorporating collection of data from multi-frequency eddy current signals. Phase analysis is used to combine the multi-frequency data to enhance the signal to noise ratio of the raw inspection image. The image is then reprocessed using a spatiotemporal filter to correlate with the frequency components of the eddy current flaw signal to separate signals associated with cracks and other flaws at edges that would ordinarily be hidden by edge effect signals.
AI classification
Ownership
GENERAL ELECTRIC COMPANY
assignment · 186930425
Assignors
SUH, UI WON, GAMBRELL, GIGI OLIVE, MCKNIGHT, WILLIAM STEWART, PISUPATI, PREETI, MISHRA, PEYUSH KUMAR, DEWANGAN, SANDEEP KUMAR, WANG, CHANGTING
On an employer assignment, the assignors are typically the inventors.