Trials on citrus fruit recognition by colour image vision

The aim of this work is to evaluate the performance of a hardware and software system for automatic recognition of the fruit present in the canopy of orange trees. This equipment carries out real time conversion from RGB to HSI colour systems. The results of various research projects favour the use of colour rather than monochromatic images and in the colour field it is the Hue-Saturation-Intensity model (HSI) that offers several advantages over the Red-Green-Blue (RGB) model. Through the system’s software, trials were also made of the identification and verification of segmentation techniques able to operate in unfavourable light conditions. Some cases of failure, although limited in number, were also examined.