CASE STUDY
Aerospace Aluminium Plates
Our client, a prominent aerospace supplier, sought our assistance in enhancing the quality control of defect inspection for their aluminum plates. Recognizing Mindtrace’s expertise in this domain, they approached us to provide a tailored solution to meet their specific requirements. We addressed this challenge by implementing a comprehensive solution that involved utilizing 3D Laser and 2D camera imaging for pipeline inspection.
To achieve this, we employed various object imaging systems, including a combination of 4 laser imaging systems and 4 cameras. The model was meticulously trained with less than 25 images for both 3D laser and 2D images, with a specific breakdown of 17 images for training and 8 for testing.
Challenges
- Minimal available training data
- Only 5 available images for testing
- Several types of pipeline defects
Image Collection
Mindrace collected all available training images, all of which were radiographic film images.
Defect Detection
The Detection Brain (D-brain) adeptly recognized diverse defects in the aluminium plates, achieving precision in anomaly detection through the use of both 3D Laser and a 2D Camera.
Outcome
Mindtrace effectively showcased the anomaly detection and complex object imaging capabilities of the Brain-Sense™ platform. Despite a limited dataset (17 training images, 8 testing images), the solution achieved an impressive accuracy of 93%. Continued image processing is expected to enhance the model’s accuracy, leveraging the continuous learning capabilities of the AI brain.
Training Images
Testing Image
Accuracy