To separate paper from cardboard you need dexterity and sharp vision. Robots make the job easier, automated, faster and more precise.
Paper and cardboard sorting is quite a challenge for traditional computer vision and is often a job for people. Our system needed to be rapidly reconfigurable to handle different materials. It needed to evolve from a static system to a dynamic system. This is where computer vision with the help of Machine learning comes in. Deep Learning offers the possibility to take into account combinations of sophisticated features such as shape, form, color and can be easily reconfigured to recognize new ‘problems’.
Within the AUTOWARE Project, imec and Robovision created a detection framework that can swiftly change between the materials. Operators can quickly train the machine to adapt. This avoids costly computer vision programming.