Unveiling the Spectrum: A Comprehensive Exploration of Carrot Varieties in India through Advanced Classification Techniques

This study introduces a new way to group carrot types into four groups: Pusa Kesar, Pusa Meghali, Pusa Yamdagni, and Pusa Rudhira. It uses an improved model that mixes a Convolutional Neural Network (CNN) with another called Random Forest (RF). Classifying carrot types is very important in farming to improve growing methods and increase the harvest. The new model uses the skills of both CNN and RF. This will make it better at classifying things because it can use pictures from this big collection of 8000 diverse types to create a combined system that uses features found by CNN while grouping smarts from RF. The CNN part gets important details from carrot photos, looking at where things are. Then the RF model makes better choices by taking many tree decisions in its calculations. In the test set, this model reached 95.7% correctness in overall results. This study helps precision farming by giving a strong way to sort out different kinds of carrots. The results show that fused models may help in investigating new mixed methods for studying and sorting crops.