Citrus cultivar identification using a computationally optimized electronic nose system and machine learning approach

Key Points

• Citrus cultivar identification achieved with a machine learning accuracy rate of over 95%.
• Key evidence from testing revealed that the system can distinguish between 10 different citrus varieties accurately.
• Implemented a computationally optimized electronic nose system to enhance scent pattern recognition during data collection.
• Technology may enable better agricultural practices; further validation in diverse environments needed.