Bitte verwenden Sie diesen Link, um diese Publikation zu zitieren, oder auf sie als Internetquelle zu verweisen: https://hdl.handle.net/10419/306193 
Erscheinungsjahr: 
2024
Quellenangabe: 
[Journal:] Amfiteatru Economic [ISSN:] 2247-9104 [Volume:] 26 [Issue:] 67 [Year:] 2024 [Pages:] 957-974
Verlag: 
The Bucharest University of Economic Studies, Bucharest
Zusammenfassung: 
This paper presents the experience and lessons learned in a pilot project aimed at integrating artificial intelligence (AI) technologies in sturgeon aquaculture. The project used convolutional neural networks and visual intelligence for the evaluation of fish biomass and the optimisation of sturgeon rearing technologies in integrated multitrophic production systems. Similar solutions have been used before to determine the biomass of other fish species, but this is the first documentation of the application of such a solution for sturgeons. The application challenges were significant, which was determined by the special morphological peculiarities of the sturgeons (shape, way of swimming, their dimensions). Both YOLACT technology and a computer vision context were tested using LAB and HSV colour spaces to estimate fish biomass based on imaging data. It was found that the LAB colour space provided superior results in terms of precision and efficiency, but maximum accuracy was achieved using convolutional neural networks (YOLACT). The analysis of the project results confirms the significant advantages of using the AI system for biomass monitoring, advantages consisting of the reduction of unit costs with labour and feed, improvement of water quality, active optimisation of sturgeon growing conditions. In this way, conditions are created for the sustainable growth of sturgeon production, both for consumption and for the restocking of various aquatic ecosystems with brood. It also proves that the large-scale implementation of AI-based technologies in the fisheries industry can make an important contribution to the achievement of Romania s National Multiannual Aquaculture Strategic Plan 2022-2030, as well as to the implementation of the European Union s strategies on food security and biodiversity.
Schlagwörter: 
artificial intelligence
deep learning
fishery industry
process optimisation
sustainability
JEL: 
Q22
Q01
O31
Persistent Identifier der Erstveröffentlichung: 
Creative-Commons-Lizenz: 
cc-by Logo
Dokumentart: 
Article

Datei(en):
Datei
Größe
863.33 kB





Publikationen in EconStor sind urheberrechtlich geschützt.