Analysis of multi-spectral images acquired by UAVs to monitor water stress of citrus orchards in sicily, Italy

Citrus production is one of the most important agricultural activities in Sicily. Consequently, it also implies significant water consumption. Considering semi-arid climate of the island, it is also prone to droughts, and some areas suffer water scarcity problems. Climate change may also exacerbate the frequency of water shortages; hence, for adaptation, it is fundamental to improve irrigation efficiency. Various technologies are emerging to this aim: installation of tensiometers and flowmeters to compare water needs and consumption so as to suggest a correction of irrigation amounts. Unmanned aerial vehicles (UAVs) mounting thermal or multi-spectral cameras may allow also a rapid assessment of the plants' stress potentially due to the need for irrigation. In this paper, we illustrate the results of preliminary research aimed at understanding whether multi-spectral images can be useful for monitoring plants' stress in citrus orchards. In particular, images were acquired through a commercial drone, the Parrot Bluegrass, mounting the Sequoia multi-spectral camera. The four-band images have been combined to obtain several vegetation indexes (VIs): the normalized difference vegetation index (NDVI), the Green NDVI (GNDVI), the leaf chlorophyll index (LCI), the normalized difference red edge (NDRE), the structure intensive pigment index (SIPI2), and the modified chlorophyll absorption in reflectance (MCARI). The maps of the indices were pre-processed in GIS to obtain single averaged values per tree. Results show that the following indices are more suitable for identifying differences in the health of orchards: NDVI, GNDVI, and SIPI2. These three indices may help identifying areas of the orchard that received a surplus or a deficit of irrigation water, or other causes of stress, finally improving orchard management and, possibly, irrigation efficiency.