One of the problems blueberry producers have is estimating production with an accurate methodology; there is generally a 5-15% error in the total volume, and the estimation value adjusts over time (from post-pruning to weeks before harvest). There are traditional methodologies but they are tedious and not very accurate.
This project was based on the creation of a system (hardware-software) for the estimation of space-time yield and quality. It had three stages: 1. Obtaining digital images at the defined control points showing the space variabilities of the analyzed plot, thus the whole study area can be extrapolated. 2. Quality estimation directly on site and subsequent lab classification, which will allow to harvest the blueberries at the optimal time of their physiological maturity. 3. Determination of the optimal time for harvest, depending on blueberry quality, from artificial vision via an autonomous capture prototype (drone).
The project managed to determine the segmentation levels according to soil texture characteristics and climate based on the plots’ spatial distribution, with satellite images on an Agroid platform. This provides control points on the site to validate the ripeness levels in order to decide on a harvest starting point.
The final results are the total kilos, beginning of harvest, and distribution of ripeness states in the different areas, according to their scattering.
| Funding institution | FIA – Hortifrut |
| Participants | FIA, INIA, Hortifrut |
| Project’s duration | March 1st 2016 to June 30th 2019 |
| Project’s total amount | $266,319,040 |
