Research Progress on Fish Disease Detection Based on Computer Vision and Deep Learning Techniques

With the rapid development of aquaculture, the frequent occurrence of fish diseases has become a major factor affecting both production yields and the quality of aquatic products. Traditional fish disease detection methods rely heavily on manual observation, which is inefficient and labor-intensive. In recent years, computer vision and deep learning-based approaches have offered new solutions for aquaculture by enabling early disease detection through the analysis of fish appearance and behavioral characteristics. This paper reviews recent advances in fish disease detection, focusing on non-contact methods based on surface abnormalities and behavioral anomalies. It explores detection principles, data acquisition methods, and the existing challenges in the field. Furthermore, corresponding solutions to these challenges are discussed. The reviewed studies highlight the technical potential of integrating computer vision and deep learning for early-stage, non-invasive, and scalable fish disease monitoring. This review also outlines promising directions for algorithm development and provides valuable references for future research and applications of computer vision and deep learning in aquatic disease detection.