Our client is a company specializing in the manufacture and sale of smart home appliances, wants to improve the quality of one of its production lines and reduce the waste produced.
In particular, they focus on the discovery of microcracks leading to leakages on gas mixer valves. If left undetected, these cracks can also break the valve during the final assembly of the device. Detecting these defects would allow a significant increase in yield, reduce cost of rework and improve the overall quality of shipped products.
However, this detection takes too long to be done manually. This is why we decided to automate it with artificial intelligence.
Sagacify first deployed its data annotation system, allowing our client's team to build a qualitative labeled dataset on which to train the model. Iteratively, a defect detection model was trained and deployed on cameras installed on the production site to automate the detection of such defects as cracks.
Key elements
The visual inspection system is in production in one offactories. It is integrated into a vision station with robotic arms that grab valves from a bin and presents them to a precision camera inspecting all surfaces. Based on these photos, the artificial intelligence algorithm makes a decision on the presence or absence of cracks.
Depending on this decision, the vision station places the valves in a valve bin for recycling, or on a conveyor line, to start production. In a cycle time of less than 3 seconds.
.png)
Feeling lost ?
Let's discuss over a coffee; our AI experts have you covered!
