 |
This is an example of how trainable vision can be employed to detect leaks. When a vessel is submerged in water, leaks can be detected by applying air pressure and looking for bubbles. Sightech's intelligent self-learning vision trains quickly to perform this difficult (for others) task. |
Step One: Train with bubbles present - sloshing water is OK. |
 |
 |
Step Two: Forget with sloshing water with bubbles absent. |
Step Three: Deploy system using Recognize mode. Note how bubbles are detected (GREEN) but the sloshing water is ignored. This demonstrates how extremely sensitive our bubble detector is. |
 |
 |
 |