Reflectance images of effective wavelengths from hyperspectral imaging for identification of Fusarium head blight-infected wheat kernels combined with a residual attention convolution neural network

Shizhuang Weng, Kaixuan Han, Zhaojie Chu, Gongqin Zhu, Cunchuan Liu, Zede Zhu, Zixi Zhang, Ling Zheng, Linsheng Huang

Abstract: Highlights•RIs of EWs from HSI were screened to analyse FHB-infected degree of wheat kernels.•RACNN was constructed for accurate identification model of FHB infection.•Optimal identification was achieved by RACNN and RIs of 940 and 678 nm.