Home
Supporting data analysis in non-destructive inspection using ultrasound.
Challenges
Ultrasonography is a safe and inexpensive non-destructive inspection method, but requires specialized knowledge and data analysis expertise, because it contains considerable amounts of noise.
Conventional AI technologies do not adequately deal with the noise inherent in ultrasound data, making it difficult to apply them to business applications.
The benefits of Ultrasound Analysis AI
A unique method specialized in ultrasound data processing facilitates high-performance analysis and expands application fields
A-mode1 signal classification
Automated classification of inspection objects using ultrasound signals
E.g.: General inspection of metals (welding, casting, etc.), concrete, etc., using a monocular probe, and inspection of frozen objects, which is generally difficult
B-mode 2 video anomaly detection
Highly accurate detection of abnormalities in the internal structure of inspection objects by training normal data only
E.g.: Inspection of food, resin packages, etc.
Features of the technology
A-mode: Improving the performance of analysis by weighting multiple ultrasound signals acquired at the same time by the size of noise
B-mode: Improved accuracy of anomaly detection by unique result display with time on the horizontal axis and normal substructure on the vertical axis
For a demonstration or to test Fujitsu Kozuchi, please get in touch