Integrating CNN and Genetic Algorithms for Kohlrabi Variety Classification in India: A Study of Purple Vienna, White Vienna, and Grand Duke

The proposed work looks into differences between types of Kohlrabi common in India using special systems called Convolutional Neural Networks (CNN) with Genetic Algorithms (GA). A set of 4000 high-quality pictures was carefully gathered and readied. These included Purple Vienna, White Vienna, and Grand Duke types. The CNN design was made to find small differences in pictures. It used convolutional, pooling, and fully connected parts for a strong grouping of things. The CNN model was good at telling apart different kinds of kohlrabi. It got it right 96.12 % of the time based on how they looked and what color they were! The test results showed that the model was very accurate across all categories, showing it is reliable in spotting small differences. Comparing it to other ways of doing things, found that the CNN model worked better. This made sure it is a strong and trustworthy choice for telling different types of Kohlrabi apart. This study shows that machine learning methods, especially CNNs can be good for farming tasks. It gives hope to finding the right type of plants better in farm work and separate markets for sales. Also, these techniques help improve breeding programs. The good use of CNNs in correctly grouping Kohlrabi types shows better ways to grow crops and can help with making food long-lasting. It also leads people closer to understanding what different plants look like, helping those who work on farms make smarter choices.