Research on the Identification of Common Economic Shellfish in Jiangsu Based on Fused-ResNet Network

Shellfish hold significant economic value and are one of the major economic crops in coastal areas of China. To identify a rapid and effective method for distinguishing different Shellfish species in Jiangsu based on morphological differences, this paper proposes a recognition algorithm for common economic Shellfish in Jiangsu based on the Fused-ResNet network. Firstly, the basic structure of the ResNet network is introduced. Then, to address the issues of similarity in shape, color, and texture among Shellfish, a Channel Spatial Attention (CSA) module is incorporated. Additionally, to more effectively differentiate the diversity of Shellfish, a transfer learning training strategy is adopted. Finally, the optimized model is applied to Shellfish recognition tasks. The proposed recognition method demonstrates universality under different targets and achieves a recognition accuracy of 98.8%. Moreover, the algorithm’s low computational cost makes it more suitable for deployment on mobile devices, providing effective technical support for the rapid identification of economic Shellfish in Jiangsu.