Giuseppe NUNNARI

Full Professor

Università degli Studi di Napoli Federico II
Country flagItaly

Research Interests

ApplyKite TwinLearns as you apply

Make your next move.

Prepare your application

Review drafts before you send them.

Explore related searches

About

Giuseppe Nunnari is a Full Professor at Università degli Studi di Catania, Italy. His research focuses on the application of deep learning techniques, including vision transformers and convolutional neural networks, for the analysis of volcanic activity. Recent publications include studies on the thermal image classification of volcanic eruptions and assessments of volcanic output rates, contributing valuable insights for hazard assessment.

Articles (7)

This paper addresses the classification of images depicting the eruptive activity of Mount Etna that were captured by a network of ground-based thermal cameras. This study aimed to evaluate the performance of Vision Transformers (ViTs), such as the Swin Transformer, compared with Convolutional Neural Networks (CNNs), including AlexNet and ShuffleNet. A dataset of 3000 images, evenly distributed across six classes, was utilized for training and testing. The results indicate that for this specific application, the performance advantage of Vision Transformers over CNNs was marginal, likely due to the nature of the classification task. While the Transformer-based models, like the Swin Transformer, demonstrated a slightly improved accuracy for certain complex classes, the CNN-based models, such as AlexNet and ShuffleNet, exhibited superior computational efficiency, particularly in terms of the classification speed. These findings highlight the suitability of CNNs for real-time volcanic activity monitoring. Additionally, this paper provides a comprehensive review of the various CNN and Vision Transformer architectures, offering insights into their strengths and limitations in the context of volcanic activity classification.

Year:

2025

This paper addresses the classification of images depicting the eruptive activity of Mount Etna, captured by a network of ground-based thermal cameras. The proposed approach utilizes Convolutional Neural Networks (CNNs), focusing on pretrained models. Eight popular pretrained neural networks underwent systematic evaluation, revealing their effectiveness in addressing the classification problem. The experimental results demonstrated that, following a retraining phase with a limited dataset, specific networks such as VGG-16 and AlexNet, achieved an impressive total accuracy of approximately 90%. Notably, VGG-16 and AlexNet emerged as practical choices, exhibiting individual class accuracies exceeding 90%. The case study emphasized the pivotal role of transfer learning, as attempts to solve the classification problem without pretrained networks resulted in unsatisfactory outcomes.

Year:

2024

Paroxysmal explosive activity at Etna volcano (Italy) has become quite frequent over the last three decades, raising concerns with the civil protection authorities due to its significant impact on the local population, infrastructures, viability and air traffic. Between 4 July and 15 August 2024, during the tourist season peak when the local population doubles, Etna volcano gave rise to a sequence of six paroxysmal explosive events from the summit crater named Voragine. This is the oldest and largest of Etna’s four summit craters and normally only produces degassing, with the previous explosive sequences occurring in December 2015 and May 2016. In this paper, we use thermal images recorded by the monitoring system maintained by the Istituto Nazionale di Geofisica e Vulcanologia, Osservatorio Etneo (INGV–OE), and an automatic procedure previously tested in order to automatically define the eruptive parameters of the six lava fountain episodes. These data allowed us to infer the eruptive processes and gain some insights on the evolution of the explosive sequences that are useful for hazard assessment. Specifically, our results lead to the hypothesis that the Voragine shallow storage has a capacity of ~12–15 Mm3, which was not completely emptied with the last two paroxysmal events. It is thus possible that one or two additional explosive paroxysmal events could occur in the future. It is noteworthy that an additional paroxysmal episode occurred at Voragine on 10 November 2024, after the submission of this paper, thus confirming our hypothesis.

Year:

2024

Paroxysmal explosive activity at Etna volcano (Italy) became very common during the last three decades, rising concern on the civil protection authorities because of its tremendous impact on the local population, infrastructures, viability and air traffic. Between 4 July and 15 August 2024, during the peak of the touristic season when the local population doubles, Etna volcano gave rise to a sequence of six paroxysmal explosive events from the summit crater named Voragine. This is the oldest and largest of the four Etna’s summit craters and normally displays just degassing, with the previous explosive sequences occurred in December 2015 and May 2016. In this paper, we use thermal images recorded by the monitoring system maintained by the Istituto Nazionale di Geofisica e Vulcanologia, Osservatorio Etneo (INGV–OE) and an automatic routine previously tested in order to automatically define the starting and ending time of each episode, the duration, total volume of erupted fluids (gas plus pyroclastic material), erupted volume of pyroclastic material, time-averaged discharge rate (TADR), and height (maximum and average) of the lava fountains. These data allowed us to infer the eruptive processes and gain some insights on the evolution of the explosive sequences useful for hazard assessment.

Year:

2024

Year:

2022

Year:

2022

Year:

2020

Collaborators (1)

Sonia Calvari

Research Director

Istituto Nazionale di Geofisica e Vulcanologia

ITALY
View details