Abstract
The aim of this work is to develop an automatic system able to provide objective information about the bloom charge in apple orchard to be used as managing tool for flower thinning activities. For this purpose was tested a mobile lab (ByeLab), equipped with: OptRx™ and LiDAR sensors (to provides vegetation index (VI) and canopy profile) an Inertial Measurement Unit (IMU) and a GNSS-RTK system. All the data were synchronized in LabView and then processed by specific algorithm in MatLab. Three surveys were carried on April 2018 during the blooming period. Preliminary investigations, using a portable spectrophotometer, were performed to better evaluate the features of the flower reflectance signature. In this way, existing and new VIs were applied and calculated to identify the better VI able to discriminate the flower to the rest of the plant. From the results obtained, the normalized difference vegetation index (NDVI) showed to be the most suitable VI to discriminate leaves from flowers; pure flowers showed NDVI values slightly negative while the leaves showed values sensitively greater than zero. Based on the result of these preliminary investigations, from the analysis of the data collected by the ByeLab a negative correlation between NDVI value and bloom charge it was expected. Despite the homogeneous behavior of the flower NDVI response, the use of OptRx™ sensors, was not able to highlight significant correlations between the number of flowers and the NDVI values. Nevertheless, merging the LIDAR and OptRx™ data it was possible to build descriptive maps that provided information about canopy thickness and VIs.