Optimized deep learning with Grad-CAM for automated cardamom classification: A multispectral imaging approach for real-time mobile deployment

Puntos clave

• Automated classification accuracy is significantly improved through optimized deep learning techniques—demonstrating a robust approach for cardamom.
• Key evidence shows that the implementation of Grad-CAM resulted in a classification accuracy of over 95% across various samples.
• Observational analysis utilizing multispectral imaging establishes a fast and efficient method for real-time classification in mobile settings.
• This work highlights potential advancements in agricultural technology, especially for field-level applications.