Context and challenges
Estimating the fruit load in orchards, during production as well as at harvest, is an important step in crop load management. This estimate is necessary in order to control the number of fruits on the tree by thinning, whether chemically, mechanically or manually. Thinning is a delicate operation to carry out as it has an impact on the current year's production, both in terms of quality (size, colour, organoleptic criteria) and quantity (number of fruits, yields), but also on production in subsequent years by limiting biennial bearing (Mathieu, 2021). By estimating the fruit load using agronomic models, it would be possible to estimate final yields and thus manage thinning and harvesting operations, storage volumes and even marketing the fruit on a larger scale.
At present, the crop load in orchards is estimated manually, either visually by observing the orchard and overall density of fruit on the trees, or by counting the number of fruits on a sample of trees chosen randomly (Vaysse, 1996), and sometimes arbitrarily, which does not always guarantee an accurate representation of variability within the orchard.