Exploring Beetroot Varieties in India: A Multi-Class Classification Approach with CNN and Random Forest Models

The study looks at three main types of beetroot - Crimson Globe, Detroit Dark Red, and Early Wonder. It uses a mixed model combining Convolutional Neural Networks (CNN) with Random Forest (RF) algorithms to classify them. A big set of 10,000 high-quality pictures taken in many different places inside India is used for this study. Before creating the model, strong cleaning methods were used first. These included standardizing and normalizing data as well as adding more examples to improve it for training purposes. The mix model uses CNN for feature getting and RF to handle data fast, aiming at top-notch classification success. The research gets a great all-around rightness of 96.2% in telling the beetroot types apart, proving that the tool is strong at finding differences between things. A careful look at how well the model works with measures like accuracy, recall and F1-score shows it does a good job of reducing wrong guesses for both classes. The talk looks at the mix-up table. It explains how good or bad this model’s sorting skills are and what could be improved in them. This study helps us learn how to sort types of beetroot. It shows that a mix between CNN and RF works well in telling different kinds apart accurately. The talk talks about ways to improve and grow the model. It highlights its possibly wide use across many different data sets and farming conditions.