UAV-Based Disease Detection in Palm Groves of Phoenix canariensis Using Machine Learning and Multispectral Imagery

Climate change and the appearance of pests pathogens are leading to disappearance palm groves Phoenix canariensis in Canary Islands. Traditional pathology diagnostic techniques resource-demanding poorly reproducible, it is necessary develop new monitoring methodologies. This study presents a tool identify individuals infected by Serenomyces phoenicis Phoenicococcus marlatti using UAV-derived multispectral images machine learning. In first step, image segmentation classification allowed us calculate relative prevalence affected leaves at an individual scale for each tree, so that we could finally use this information with labelled situ data build probabilistic model detect specimens. Both pixel performance model’s fitness were evaluated different metrics such as omission commission errors, accuracy, precision, recall, F1-score. It worth noting accuracy more than 0.96 obtained healthy leaves, good detection ability model, which reached 0.87 trees. The proposed methodology presented efficient identifying specimens, spectral information, reducing need fieldwork facilitating phytosanitary treatment.