RDB-CSCA: A deep learning method to identify the spectral information for tracing the egg quality

Abstract

It is important to provide a fast and nondestructive detection method for egg quality supervision to realize the quality identification of eggs in different storage periods based on sensing detection technology. In this paper, a detection method is proposed for egg quality identification in different storage periods based on a deep learning method combined with a hyperspectral system. First, channel-spatial collaborative attention (CSCA) is proposed to adaptively focus on the critical features of deep spectral information. Second, residual dense block (RDB) network is introduced to adaptively fuse deep and shallow features, thereby avoiding the feature degradation phenomenon that occurs in the convolution calculation process. Finally, the CSCA and RDB are combined to construct a decision-making network for the identification of spectral information of eggs in different storage periods. The results show that in the comparison of multi-structure models, RDB-CSCA achieves the best classification performance, with the classification accuracy of 96.88%, precision of 97.18%, recall of 97.61%, and F1 score of 97.39%.