The Essential oil (EO) extracted from bergamot skin (Citrus bergamia, Risso et Poiteau) is highly appreciated for its use in perfumery and gastronomy. Notably, 90% of the EO bergamot production is concentrated in the Province of Reggio Calabria (Southern Italy) under a protected designation of origin (PDO). Early yield estimation of the essential oil content of fruits is fundamental to helping growers make farming decisions, especially regarding harvesting operations. The application of advanced modelling techniques based on artificial intelligence and digital device technology can contribute to this goal. This thesis represents a three-year study dedicated to the development of an Android application aimed at determining the essential oil content in Bergamot fruits. The entire process is divided into three distinctive phases, which cover fruit sampling and oil extraction, the use of deep learning techniques, in particular convolutional neural networks (CNN), and the practical implementation of the Android application. The first phase involved the systematic sampling of Bergamot fruits and the extraction of the essential oil, establishing the basis for the subsequent analysis. In the second phase, advanced deep learning techniques, mainly CNNs, were used to process the collected data. This phase involved the development of a robust algorithm capable of accurately predicting essential oil content based solely on fruit color. The third and final year was dedicated to the practical implementation of this algorithm within an Android application. Using Android Studio, an accessible and easy-to-use application was created, allowing users to quickly and conveniently exploit the results obtained from the data analysis phase. The methodological approach adopted in this work not only offers an effective means to determine the essential oil content in Bergamot fruits through color recognition, but also represents a tangible demonstration of the practical application of deep learning techniques in the agricultural sector, together with the development of advanced applications for Android devices. This study not only contributes to the growing body of knowledge in the field of artificial intelligence and agriculture, but also constitutes a significant step towards the convergence of these disciplines, paving the way for new perspectives in agricultural management and simplified access to solutions innovative technologies.
L'Olio Essenziale (EO) estratto dalla buccia di bergamotto (Citrus bergamia, Risso et Poiteau) è molto apprezzato per il suo utilizzo in profumeria e gastronomia. In particolare, il 90% della produzione di bergamotto OE è concentrato nella provincia di Reggio Calabria (Italia meridionale) sotto una denominazione di origine protetta (DOP). La stima anticipata della resa del contenuto di olio essenziale dei frutti è fondamentale per aiutare i coltivatori a prendere decisioni agricole, in particolare per quanto riguarda le operazioni di raccolta. L’applicazione di tecniche di modellazione avanzate basate sull’intelligenza artificiale e sulla tecnologia dei dispositivi digitali può contribuire a questo obiettivo. Questa tesi rappresenta uno studio triennale dedicato allo sviluppo di un'applicazione Android volta a determinare il contenuto di olio essenziale nei frutti di Bergamotto. L'intero processo è suddiviso in tre fasi distinte, che riguardano il campionamento dei frutti e l'estrazione dell'olio, l'utilizzo di tecniche di deep learning, in particolare le reti neurali convoluzionali (CNN), e l'implementazione pratica dell'applicazione Android. La prima fase ha previsto il campionamento sistematico dei frutti di Bergamotto e l'estrazione dell'olio essenziale, ponendo le basi per le successive analisi. Nella seconda fase sono state utilizzate tecniche avanzate di deep learning, principalmente CNN, per elaborare i dati raccolti. Questa fase ha comportato lo sviluppo di un robusto algoritmo in grado di prevedere con precisione il contenuto di olio essenziale basandosi esclusivamente sul colore del frutto. Il terzo ed ultimo anno è stato dedicato all'implementazione pratica di questo algoritmo all'interno di un'applicazione Android. Utilizzando Android Studio è stata creata un'applicazione accessibile e facile da usare, che consente agli utenti di sfruttare in modo rapido e comodo i risultati ottenuti dalla fase di analisi dei dati. L'approccio metodologico adottato in questo lavoro non solo offre un mezzo efficace per determinare il contenuto di olio essenziale nei frutti di Bergamotto attraverso il riconoscimento del colore, ma rappresenta anche una dimostrazione tangibile dell'applicazione pratica di tecniche di deep learning nel settore agricolo, insieme allo sviluppo di applicazioni avanzate per dispositivi Android. Questo studio non solo contribuisce al crescente corpus di conoscenze nel campo dell’intelligenza artificiale e dell’agricoltura, ma costituisce anche un passo significativo verso la convergenza di queste discipline, aprendo la strada a nuove prospettive nella gestione agricola e all’accesso semplificato a soluzioni tecnologiche innovative.
Development of a mobile web application to assess the essential oil content of bergamot in the field from smartphone images using convolutional neural networks / Anello, Matteo. - (2024 Sep 05).
Development of a mobile web application to assess the essential oil content of bergamot in the field from smartphone images using convolutional neural networks
Anello, Matteo
2024-09-05
Abstract
The Essential oil (EO) extracted from bergamot skin (Citrus bergamia, Risso et Poiteau) is highly appreciated for its use in perfumery and gastronomy. Notably, 90% of the EO bergamot production is concentrated in the Province of Reggio Calabria (Southern Italy) under a protected designation of origin (PDO). Early yield estimation of the essential oil content of fruits is fundamental to helping growers make farming decisions, especially regarding harvesting operations. The application of advanced modelling techniques based on artificial intelligence and digital device technology can contribute to this goal. This thesis represents a three-year study dedicated to the development of an Android application aimed at determining the essential oil content in Bergamot fruits. The entire process is divided into three distinctive phases, which cover fruit sampling and oil extraction, the use of deep learning techniques, in particular convolutional neural networks (CNN), and the practical implementation of the Android application. The first phase involved the systematic sampling of Bergamot fruits and the extraction of the essential oil, establishing the basis for the subsequent analysis. In the second phase, advanced deep learning techniques, mainly CNNs, were used to process the collected data. This phase involved the development of a robust algorithm capable of accurately predicting essential oil content based solely on fruit color. The third and final year was dedicated to the practical implementation of this algorithm within an Android application. Using Android Studio, an accessible and easy-to-use application was created, allowing users to quickly and conveniently exploit the results obtained from the data analysis phase. The methodological approach adopted in this work not only offers an effective means to determine the essential oil content in Bergamot fruits through color recognition, but also represents a tangible demonstration of the practical application of deep learning techniques in the agricultural sector, together with the development of advanced applications for Android devices. This study not only contributes to the growing body of knowledge in the field of artificial intelligence and agriculture, but also constitutes a significant step towards the convergence of these disciplines, paving the way for new perspectives in agricultural management and simplified access to solutions innovative technologies.File | Dimensione | Formato | |
---|---|---|---|
Anello Matteo.pdf
accesso aperto
Tipologia:
Tesi di dottorato
Licenza:
DRM non definito
Dimensione
4.05 MB
Formato
Adobe PDF
|
4.05 MB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.