Unveiling Kale Diversity in Indian Agriculture: CNN - Logistic Regression Classification

The presented study looks into the different types of kale used in India by using Convolutional Neural Networks (CNN) and Logistic Regression. collected around 12.,000 high-quality pictures of three types of kale - Dwarf Green Curled., Siberian., and Red Russian. made sure these images were ready to use before starting working with them. The information was split into 70% to train the model, 15% for checking it out, and another 15% to test. The suggested design includes layers for feature getting and Logistic Regression to make final decisions. The model is 95.18% accurate and can strongly tell the difference between different types of kale. use accuracy, completion rate, and F1-score to check if its strong performance is the same for all classes. This guarantees good classification every time. This study shows how machine learning methods can help with farming tasks, giving a good automatic system for sorting different kale types. Future work could be about making the model better for big things, faster computers, and using it properly in real farms. This will help grow crops that are good for everyone in India while using resources well.