Automation of agricultural processes is one way to deal with the expected increase in food demand. In this work, the CAL is challenged to study how to introduce a manipulator into a tomato transplanting process.  The aim is to reduce the number of labor by introducing manipulators to the widespread semi-automatic transplanters.

The main problem that must be solved in introducing a manipulator in a semi-automatic transplant is the automatic detection of the plant positions.  First, it is necessary to estimate the box where plants lie. Once that the box position is known, the problem is reduced into a pick-and-place task. With the estimated box position, the stem position can be roughly estimated, since the distance between the holes and
their total number are known.

The proposed laboratory setup consisting of a manipulator, a vision subsystem and ROS shows the feasibility of having a partially automated tomato transplanting machine using a manipulator (Figure 1, Figure 2).

Figure 1: The proposed system.


Figure 2:  Vision subsystem scheme.



Figure 3: Position measurement of the gripper with a grabbed plant. The plant thickness average is 2.85 mm.