Quantifying the Production of Fruit-Bearing Trees Using Image Processing Techniques

Abstract

[EN] In recent years, the growth rate of world agricultural production and crop yields have decreased. Crop irrigation becomes essential in very dry areas and where rainfall is scarce, as in Egypt. Persimmon needs low humidity to obtain an optimal crop. This article proposes the monitoring of its performance, in order to regulate the amount of water needed for each tree at any time. In our work we present a technique that consists of obtaining images of some of the trees with fruit, which are subsequently treated, to obtain reliable harvest data. This technique allows us to have control and predictions of the harvest. Also, we present the results obtained in a first trial, through which we demonstrate the feasibility of using the system to meet the objectives set. We use 5 different trees in our experiment. Their fruit production is different (between 20 and 47kg of fruit). The correlation coefficient of the obtained regression model is 0.97.