Development of an implement fitted with sensors for massive data capturing in vineyards

One of objectives of NOVATERRA project is the development of a system to carry out variable rate application of plant protection products based on canopy maps made from satellite images. In addition, these satellite maps must be validated in the field to develop models that allow the characterization of the vegetation in vineyards and olive groves to be estimated automatically from satellite images. For this, in the first year of NOVATERRA, an implement has been developed with square steel profiles that support 4 ultrasound sensors, a vision system consisting of 2 cameras and a GNSS geopositioning system. The electronic control unit developed is based on Raspberry Pi interface. This implement is suspended from the rear of the tractor and with a frequency of one second it calculates the width and height of the vegetation, and also captures an imatge from the canopy with its cameras in such a way that a detailed information for canopy characterization (volume and density) is available. With a forward speed of up to 6 km/h, the system outputs a map of points (positions of each reading) and all the variables recorded for that point.

This is an improvement over the current way of establishing these measurements as it considerably reduces the time required to manually measure in the field. In addition, it allows establishing vegetation volume maps (m3/ha) that translate directly into the volume of liquid to be applied (l/ha). On the other hand, this way of measuring allows massive data generation to validate the relationships between the satellite data and the real measurements of the vegetation, allowing to work with more points within the plot, and making the models developed in future tasks of NOVATERRA more reliable and robust.

Author