Denmark is the country in Europe which annually produces the most festive greenery, Christmas trees and related products. The cutting and processing of the plant material in this area is still mostly done by manual labor and carried out under unfavorable working conditions. Pyntegrønt tries to change this with the help of computer vision and robotics.
The project is based on a combination of camera and robot technology, which will help to color sort the collected tree branches. A technology that can help to keep large parts of the production in Denmark, which will in turn facilitate the physical workload significantly.
SDU Robotics deals in this project with the classification of the individual branch parts into multiple classes (stem, blue category, green category, moss, algae, …) and based on that the determination of cutting locations to get the best economic return. We use line scan technology to get high quality images which are then used in a deep learning approach for classification of parts.
Official project title:
Aut. vision- og robotteknologi til bæredygtig og konkurrencedygtig produktion af pyntegrøntprodukter
Contact:
Associate Professor Dirk Kraft
Project participants:
WEFRI A/S; KP-Automation A/S; SDU Robotics
Funding agency and instrument:
NaturErhverstyrelsen, GUDP