Please login first
Probabilistic classification of infected palm trees using UAV-based multispectral imagery and machine learning
* 1 , 2 , 3 , 2 , 1
1  Departamento de Física, Universidad de La Laguna, 38200 San Cristóbal de La Laguna, Spain
2  Departamento de Biología Animal, Edafología y Geología, Universidad de La Laguna, 38200 San Cristóbal de la laguna, Spain
3  Sección de Sanidad Vegetal, Dirección General de Agricultura, Consejería de Agricultura, Ganadería y Pesca, 47014 Santa Cruz de Tenerife, Spain
Academic Editor: Giorgos Mallinis

Abstract:

Human activities have led to the global redistribution of species, causing a worldwide decline in biodiversity, followed by the apparition of non-native species in natural environments, jeopardizing the normal function of the ecosystems and the apparition of invasive pests and new pathogens to crops and forests. The Canarian archipelago landscape, shaped by Canarian palm tree (Phoenix canariensis) groves, is being affected by this phenomenon, with their consequent decline.

The European Union Natura 2000 protection areas designated Phoenix canariensis groves as a priority habitat as an essential endemic Canary Islands plant species, and in this context, new tools to monitor and treat the pathologies that affect this species.

Traditional pathology diagnostic techniques are resource-demanding and poorly reproducible, and it is necessary to develop new monitoring methodologies. This study presents a tool to identify individuals infected by Serenomyces phoenicis and Phoenicococcus marlatti using UAV-derived multispectral images, machine learning, and probabilistic classification techniques. Two different study areas were selected in Tenerife and La Gomera islands, due to their representativity of the health status of Canarian palm groves.

In the first step, image segmentation was used to automatically identify palm tree specimens. In the second step, a pixel-based classification allowed us to assess the relative prevalence of affected leaves at an individual scale for each palm tree. The calculated affection prevalence ratio was later used alongside labelled in situ data depicting healthy and infected individuals, collected by expert technicians’ visual inspection to build a probabilistic classification model, capable of detecting infected specimens. Both the pixel classification performance and the model’s fitness were evaluated using different metrics such as omission and commission errors, accuracy, precision, recall, and F1-score.

An accuracy of more than 0.96 was obtained for the pixel classification of the affected and healthy leaves, and the probabilistic classification model presented good detection ability, reaching an accuracy of 0.87 for infected palm trees.

It is worth considering that the developed algorithms and the infection detection model could allow for the cost-effective identification of infected palm trees by implementing transfer learning procedures in new study areas. This will imply a drastic decrease in data requirements, facilitating future palm groves' extensive monitoring in the archipelago, and significantly reducing phytosanitary treatment costs.

Keywords: probabilistic classification modelling; support vector machine; random forest; spectral separability analysis; structure insensitive pigment index; NDVI; Canary Islands
Comments on this paper
kalyl cie
Traditional pathology diagnostic techniques are resource-intensive and often lack reproducibility. In response to these challenges, there is a growing need for innovative monitoring methodologies to assess and treat the health of Canarian palm groves.

Nico England
I have read your paper and I found your paper very informative.

Michael Saldana
I think Unmanned Aerial Vehicles (UAVs) equipped with multispectral sensors are used to capture high-resolution imagery of palm tree plantations. Multispectral sensors capture data across different wavelengths, providing detailed information about the health and condition of palm trees. The acquired UAV imagery is preprocessed to remove noise, correct for atmospheric effects, and align the images properly. This ensures the accuracy and consistency of the data for subsequent analysis. Relevant features are extracted from the multispectral imagery to characterize the health status of palm trees. These features may include vegetation indices (e.g., NDVI, NDWI), texture measures, and other spectral properties that can differentiate between healthy and infected palm trees.

kalyl cie
Two distinct study areas in Tenerife and La Gomera islands were selected for their representativeness of Canarian palm grove health. The methodology involves image segmentation to automatically identify palm tree specimens, followed by pixel-based classification to assess the relative prevalence of affected leaves at an individual scale. A probabilistic classification model, trained using labeled in situ data collected by expert technicians, is then utilized to detect infected specimens.

Kathy Barrera
I believe that multispectral sensor-equipped Unmanned Aerial Vehicles (UAVs) are utilized to obtain high-resolution aerial photos of palm tree plantations. With the use of multispectral sensors, which collect data at many wavelengths, palm tree health and condition may be thoroughly examined.



 
 
Top