On behalf of Mettler Toledo (Albstadt) GmbH a vision based system was developed at Fraunhofer IITB for automatically detecting and recognizing fruits and vegetables. The system is currently in use in German super markets and consists of a monocular color camera and a scale.
Many fruits have different colors and textures due to different degrees of maturity, e.g. banana. Other fruits, like apples or pears, provide several types with different colors and shapes. Finally, often the fruits are packed into plastic bags which mostly give the fruits a totally different texture due to reflexions and shadows.
Banana with different degrees of maturity and with two differnt color cameras.
Differend kinds of pears and apples.
Besides the solution of the above described challenges, the IOSB solution is robust against color and brightness deviations, which makes the system applicable in a variety of locations and scenarios. Furthermore, the system is very user friendly and intuitive. Employees in the local stores are able (without any specific knowledge) to train the system for new fruits - quick and easy.