Deep Learning algorithms teach robots and machines to do what is natural for human beings: learn by example. They are based on how neural networks operate. They try to simulate the functioning of the human brain through learning and image association by example.

Deep Learning-based algorithms also learn from data and examples and are able to draw their own conclusions. They learn hierarchically, in layers. The more layers there are, the more complex the algorithms are, allowing for better conclusions to be drawn.

Advantages of Deep Learning:

— It improves complex image recognition.

— It allows for rapid adaptation to new models without the need for complex programming changes.

— It tolerates deviation and defects that are unpredictable for conventional vision systems.

— It combines the flexibility and application of logical criteria of human vision with the safety and speed of a vision system.

Want more information? Contact us - we'll be happy to help you!

Bin Picking systems are vision-guided robotic (VGR) solutions that automatically select and extract stacked parts, in a disordered environment, randomly from a container or box.

Our solution uses a system of vision of recognition and localization of pieces as well as a robotic system for extraction and unloading in the later work station, by means of a specific software for the treatment of the image and the calculation of trajectories, avoiding collisions.

At teknics we offer our clients the ease of having a single interlocutor throughout the project, from its simulation, feasibility study, prototype, design, construction and commissioning, offering turnkey solutions.