Multimodal fusion-based classification method for small-sample imperfect wheat kernels using hyperspectral imaging

Key Points

• The classification method accurately identifies imperfect wheat kernels, potentially enhancing quality assessment.
• Key evidence shows an accuracy of over 90% in classification using hyperspectral imaging and fusion techniques.
• Analysis using multimodal fusion-based techniques integrates data from various sources for improved outcomes.
• Highlights the need for effective classification methods in agricultural practices to ensure better quality control.