Artificial Vision Based Smart Solution for Blueberry Harvest Automation

The technological goal of this experiment is to develop a data processing pipeline to detect, classify and count blueberries and to measure maturity automatically using a Deep Learning image segmentation method for blueberry cultivars that can be integrated in a smart solution. A prototype connected to a camera in one of the electric tractors developed by AGRIA ZE will be deployed to be tested as a Proof of Concept of the technologies and their digitalization capacity. Blueberry's main quality indicators associated to consumer acceptability are related to the fruit appearance and texture. Today, appearance quality assessment is determined subjectively using visual observation. Computer vision analysis is a useful tool to evaluate fruit optical properties. An accurate and reliable image-based quantification system for blueberries additionally is useful for the automation of the complete harvest process management. It also serves as the basis for controlling robotic harvesting systems. Quantification of blueberries from images is a challenging task due to occlusions, differences in size, illumination conditions and the irregular amount of blueberries that can be present in an image. In blueberries, there is a large variation in fruit maturation among clusters (fruit within a cluster), and among branches of the bush. This means that mature fruit must be harvested selectively after each fruit develop a blue colour. Therefore, evaluating the maturity of individual blueberry fruit can provide valuable information for determining the best harvesting time. Number counting and cluster tightness are also important in the blueberry industry and can be used to estimate yield and harvesting strategy. Thus, an automated method to characterize these three traits (compactness, maturity, and count) and present the analysis would be a valuable tool for blueberry breeders as well for blueberry growers. Thus, the main technical challenges to answer sector demands are: