Project by: Joost Kingma, Sander van Gemmert, Ka Chun Tsang and Mark van Leeuwen
Hoogeveen Plants is a horticulture company that specialize in Climbers, Fruit Plants, Bamboo-Grasses-Ferns and Helleborus. Hoogeveen Plants is looking for innovative ways to automate their process. Their current project is to automate the binding process of the Climbers. In the current situation the Climbers are bound manually, sorted by the growth stage.
The main goal of our project is to design a machine that can bind the Climbers automatically. To reach this goal we discussed the binding and the growth process of the Climbers.
The requirements for this project:
• The Climbers can’t be damaged
• The dimensions of the pot and the wooden frame can’t be changed
• Separate the Climbers for binding
Our end system is divided in multiple subsystems consisting of a conveyor belt, universal robot with an end of arm tool, PLC, HMI and a Realsense camera (SR300).
End of arm tool
We got a plant binder from the client, which is used by the employees to bind the Climbers manually. To attach the binder to the robot we had to disassembled it and remodify it for the robot.
To detect the Climbers with vision we programmed in Python with the OpenCV library. OpenCV is an open source library mainly aimed at computer vision. The vision consists of two major parts, detecting the wooden frame and the branches.
The first part of the vision is detecting the frame. By detecting the frame, we can get the following data: the pixel points where we can bind. The frame can be split in 3 parts (lowest, middle and highest) and will be used in the next part.
The second part of the vision is detecting the Climbers. The vision algorithms will detect if it’s bonded in the lower, middle and higher part of the frame. It will show the coordinates where to split the Climbers and where to bind it.
The PLC drives this system, it is connected to the universal robot, HMI (Human Machine Interface) and the laptop running the vision program. The HMI sends signals to the PLC to start the program or to manually control each part of the system.
The main goal of this project is to bind the Climbers. To create a process which can detect the branches and bind it without damaging it. We have succeeded that goal for climbers in winter rest.