Deep learning for machine vision, identification and quality control
Artificial intelligence and especially deep learning have recently enabled automation of tasks that requires coping with huge variation in circumstances and / or products.
Fizyr, previously Delft Robotics, serves many multinationals with applying 3D vision for 6DOF manipulation, order picking and parcel handling. The same deep learning algorithms also perform robust identification and accurate quality control.
In this presentation we will discuss recently achieved projects such as creating a picking robot next to an AutoStore which included identifying and picking the right items from a bin.
For PostNL we have made a robot picking unknown postal parcels from a roll cage. Last but not least, we will also show recent cases of applying deep learning for identifying and quality control in agri, food and horticulture.