I&T Solution |
Hauling Rope Flaw Inspection System
(REF: S-1267) |
Trial Project |
|
Solution Feature |
- Computer Vision Rope Flaws Detection (CVRFD) method is using computer vision deep learning technology.
- CVRFD consists of three main parts, which are data collection, model training, and inference.
- The quality of photos is crucial for CVRFD, traditional image preprocessing techniques are also used to increase image quality.
- FPN (Feature Pyramid Network) with ResNet101 backbone is recommended
|
Trial Application and Expected Outcome |
- Malformation can be detected by structure and texture features.
- Broken can be detected by the surface texture features
- Corrosion can be detected by the brown and yellow color features.
- Abrasion can be detected by the flat surface texture features
- Fatigue can be detected by the structure features
|
Additional Solution Information |
Hauling_Rope_Flaw_Inspection.pdf
|
Info on I&T Solution Provider |
|