Automatic identification of unharvested table olives in hyperspectral imaging for decision-support applications

Abstract

During the growing season, olives progress through nine different phenological stages, starting with bud development and ending with senescence. Determining the appropriate moment for collecting olives depends on their intended purpose, and the first step in this process is identifying the specific stage of phenological development. In this study, our objective is to develop a tool capable of identifying olives through their spectral signature. To achieve this goal, we utilize hyperspectral imaging of olives while they are still on the tree and conduct this process throughout the entire growing season, from May to September, in the field. Images were taken in the field every week during the table olive season from 9:00 to 11:00. To analyze the data, we trained and tested classifiers including Decision Tree, Logistic Regression, Random Forest, and Support Vector Machine on labeled datasets. The Logistic Regression model demonstrated the best performance in olive identification and performance metrics. It is possible to identify olives in a hyperspectral image throughout the season using this novel model. In subsequent studies, this data can be utilized to develop practical applications to enhance a farmer's decision-making process. In doing so, producers can evaluate production metrics using a non-invasive technique.