A Ground-based Platform for Reliable Estimates of Fruit Number, Size, and Color in Stone Fruit Orchards

Automatic in-field fruit recognition techniques can be used to estimate number, size, skin color, and yield in crops. Fruit color size represent two of the most important quality parameters stone ( Prunus sp.). This study aimed evaluate reliability a commercial mobile platform, sensors, artificial intelligence software system for fast estimates peach persica ), nectarine P. var. nucipersica plum salicina apricot armeniaca assess their spatial temporal variability. An initial calibration was needed obtain absolute number per tree forecasted yield. However, technology also produce relative density maps orchards. prediction accuracy ?90% all crops training systems under study. Overall, predictions two-dimensional were slightly more accurate. Estimates diameter (FD) did not need an calibration. The FD had percent standard errors <10% root mean square error <5 mm different systems, row spacing, crops, position within canopy. Hue angle, attribute previously associated with maturity nectarine, that best predicted by platform. A new parameter—color development index (CDI), ranging from 0 1—was derived hue angle. adoption CDI, which represents progression or distance green, improved interpretation measurements end-users as opposed angle generated robust estimations turn purple when ripe, such dark plum. Spatial FD, CDI obtained platform inform orchard decisions thinning, pruning, spraying, harvest timing. importance application crop real-time assessments forecasts are discussed.