ChallengeRocket

NVIDIA® Jetson™ Developer Challenge

CURRENT STATUS

Winners announced

SEE RESULTS

Thanks to all Participants!

SEE RESULTS

NVIDIA® Jetson™ Developer Challenge

Back

Cognitive scanner for quality inspection

Project description

We have created a cognitive scanner for quality inspection. The solution is built around the Nvidia Tx1 devkit with 2 cameras connected to cover as much as possible from the produced item.

Our cameras has builtin led light to make the classification as consistent as possible. The cameras are easy to adjust by using flexible arms.

The cameras are easy to connect to by Wifi. HDMI and USB ports are available at the side of the cabinet.
Two push buttons are available to make it easy to start and stop.

The cameras are filming in 30 FPS in a resolution of 12 MP. This is possible by our custom software stack. Each camera is running its own inference engine, the classification result is then sent to our platform called Skyraid.

The inference engine is built with Cuda and TensorRT for maximum performance.

The Skyraid platform is built to be the control plane of the solution. It receives the events from each camera and then routes the events to the correct configured application. The application can then make the decision if a relay ouput should be set or just a email should be sent.

Our team consist of

Marcus, Programming / Electrical Engineering
Pawel, Mechanical / Electrical Engineering
Peter, Business Developer
John, Customer with domain knowledge
Henrik, Photo and movie