Study on the Flow Velocity Preference of the Four Major Chinese Carps Using Convolutional Neural Networks

Flow velocity is a critical factor in determining the suitability of fish habitats. Understanding the preference patterns of the four major Chinese carps (FMCCs) for different flow velocities is crucial for their habitat conservation and restoration. In this study, the preference of individual fish species, approximately 15 cm in length, for flow velocity was investigated at flow velocity gradients of 0.0, 0.4, 0.8, 1.2, 1.6, and 2.0 times their body length. Additionally, a deep learning algorithm based on convolutional neural networks (CNNs) was employed for fish target detection. The results showed that, at this length, black carp (Mylopharyngodon piceus) preferred fast currents when the inlet flow velocity was between 0.4 and 1.6 times their body length, while grass carp (Ctenopharyngodon idella), silver carp (Hypophthalmichthys molitrix), and bighead carp (Hypophthalmichthys nobilis) preferred fast currents when the inlet flow velocity of the test flume was between 0.4 and 2.0 times their body length. However, this preference for fast currents decreased as the overall flow velocity increased to a specific threshold, eventually leading to their avoidance. The highest preference for fast currents among the four species was observed at inlet flow velocities of 1.2, 0.4, 0.8, and 0.8 times their body length, respectively. The findings of this study provide important insights into habitat conservation and restoration for the FMCCs in projects focused on the construction of navigation channels and water conservancy.