robot for cucumber was also developed, based on the cucumber
picking robot as a platform (Van Hentenet al., 2006)byusingan
information-based design method (Van Tuijlet al., 2004). The
de-leafing robot did not have any leaf collection function but
dropped the leaves onto the ground after cutting a leaf stalk
using an end-effector. The time required for the de-leafing
robot to detect the leaf and cut the leaf stalk was approximately
140 s per leaf. Consequently, if efficiency is to be increased, an
improvement in the automated leaf picking process is required.
The desired time for the automated leaf picking should be in
thesameorderofmagnitudeasthetimeforthemanual
picking time, which is 5–10 s per leaf.
Following preliminary experiments in a Japanese greenhouse (Otaet al., 2005), a prototype of a cucumber leaf picking
device was designed and evaluated in Dutch greenhouses
using an adapted vehicle. The objectives of this research
were (1) to develop a laboratory experimental apparatus for
the leaf picking device; (2) to conduct picking tests in the
laboratory to evaluate its performance; (3) to analyse the
relationship between mechanical conditions and the removal
performance; (4) to conduct a greenhouse experiment with
cucumbers grown with the pinching training system and to
investigate performance under practical conditions.
2. Materials and methods
2.1. Laboratory experimental apparatus
The picking rotor was composed of stainless-steel knives
and/or brushes made from stainless steel wire (Fig. 1). The
picking rotor was designed and manufactured in such a way
that the configuration of knives and brushes could be
changed. Knives and brushes were alternately spaced at
equal distances around the circumference of the rotor at 901
intervals. Leaves were extracted into the housing by the
pulling action of the rotating rotor and suction from the
vacuum cleaner. Leaves were shredded by the rotor in
the housing.