Cambridge-based Sound Mathematics Ltd have created ‘ImageTriage,’ an extensible software platform for interpreting disorganised and haphazard images.
Capable of automatically processing non-destructive testing (NDT) and structural health monitoring (SHM) data and creating draft inspection reports, ImageTriage employs decision trees and fuzzy logic rather than deep learning AI, meaning that it is data lean, agile, explanatory and energy efficient.
The software system employs signal processing functions for surface profiling, meshing, ray tracing and intensity generation to create two-dimensional total focusing method (TFM) images, sectorial images, a series of images using different portions of the phased array ultrasonic testing (PAUT) results, or a combination of scans created by ordinary probes. The images are processed with fast algorithms from the open source computer vision library (OpenCV).
The decision trees will then select those images that seem to contain defects, evaluates their characteristics, groups similar images together and delivers a draft report.
The software offers potential solutions in the NDT / SHM market as, despite being originally designed for PAUT, it can inspect any inspection data delivered as images, supporting a wide range of inspection scenarios without major redesign.
Find Out More:
With a number of potential space sector applications for this system, you can find out more about this technology here:
If you would like to explore this technology further for potential applications, please get in touch with us by emailing contactus@esa-technology-broker.co.uk.