Machine vision enables greenhouse automation

Using two inspection stations based on smart cameras, a Dutch horticultural company checks the size and health of coconut palms, automating labor-intensive tasks.

Oct 16th, 2008

To improve efficiency and reduce labor costs in its greenhouses, a Dutch horticultural company installed inspection stations based on smart cameras to check the size and health of coconut palms. The company, Kwekerij G. Verkade (Honselersdijk, The Netherlands) grows coconut palms in greenhouses that cover a floor space of 40,000 m2.

The coconuts are potted manually and stored in a dark, warm hall with a high humidity level. After germinating, the plants are moved to other greenhouses where they reach the desired size. Separating healthy plants from diseased or undersized ones, sorting out coconuts that have not sprouted, and measuring plants for sale has been labor-intensive and time-consuming. The company therefore aimed at automating these tasks.

Verkade turn to Sedeco Vision Components (Mijdrecht, The Nether-lands; www.sedeco.nl), which specializes in image-processing systems, to design a testing facility with two stations based on VC2065EC cameras from Vision Components (Ettlingen Germany; www.vision-components.com). The potted plants are picked up by a forklift and separated on a conveyor belt that leads past the inspection stations. Depending on the quality of the coconut palms, the cameras send a signal to a PLC that then directs the plants to an intermediate storage facility or to a waste container via conveyor belts.

The first inspection station serves to examine seedlings--viable plants are transported from the intermediate storage to a greenhouse; all others are rejected. During their growth, the coconut palms are regularly examined at the second inspection station. There, they are sorted into four groups according to size, and each group is transported to a different greenhouse. Diseased palms are identified by the color of their leaves and then rejected, as are plants that no longer grow compared to their group.

The fully automated inspection facility allows users to quickly check whether a coconut has sprouted and whether the plant is healthy. Unsuitable specimens are sorted out as soon as possible, which optimizes the use of floor space. Moreover, dividing the plants into different size groups makes it unnecessary to measure plants for sale by hand. In this way, product quality is improved and personnel costs are minimized.

More in Environment & Agriculture