Abstract

The challenge of the modern fruit-growing industry is to increase yields while maintaining quality. The methodology proposed in this research could help to meet this challenge by using a multispectral and RGB digital images, for estimating weighted average (g), accumulated production (kg), and soluble solids (°Brix) in apples, a Royal Gala cultivar. The method consists of segmenting the canopies of trees and dividing them by the Normalized Difference Vegetation Index (NDVI) to obtain a Radiometric Foliar Index (RFI). The number of apples per tree is then normalized by the IFR in order to obtain a Radiometric Productive Foliar Index (RPFI). The RPFI was evaluated in the 2006/2007 season in a cv “Royal Gala” apple orchard. A multi-spectral image of the studied area was taken, from which the NDVI vegetation index was obtained, classifying this in 3 classes or zones of vigor: high (HV), medium (MV) and low (LV). Three levels of crop load were established in each of the sites: low or normal load of the orchard (1-3 fruits per cm2 of Trunk Cross Sectional Area, TCSA), medium with 5-7 fruits per cm2 of TCSA and high with 7-10 fruits per cm2 of TCSA. The fruit weight was quantified at harvest (PF, in g.) for 100% of the sample, that is, for more than18,500 apples and sub-samples of 20 apples per experimental unit and block in order to obtain °Brix. The RPFI explained 84%, 60% and 51% of the variance for production per tree (kg), weighted average (g) and soluble solids (°Brix), respectively, being the best model in comparison to the crop load expressed in the Trunk Cross Sectional Area (TCSA; r2 = 0.47, 0.48, and 0.41, respectively), canopy volume (CV; r2 = 0.31, 0.31, and 0.29, respectively) and canopy area (AC ; r2 = 0.75, 0.57, and 0.44, respectively).