From Field to Algorithm: A New Strategy for Pusa Gourd Classification Using Advanced CNN and RNN Models in India

This study examines the classification of several types of Bottle Gourds in India using a novel hybrid model of Convolutional Neural Networks (CNN) and Recurrent Neural Networks (RNN). A collection of 8200 images depicting four primary varieties of bottle gourds - Pusa Samridhi, Pusa Naveen, Pusa Meghdoot, and Hybrid 1 from the “Pusa” group - was meticulously assembled and processed before utilization. The strategy for constructing it, by integrating a CNN specialized in image analysis and a RNN proficient in sequential processing, achieved remarkable results with an accuracy rate of 97.4%. Upon rigorous examination, the model demonstrated exceptional proficiency in discerning the minute distinctions that distinguish each variety of plant. Performance metrics like as accuracy in making correct predictions, ability to identify all correct answers even when they were difficult to discern, and a balanced combination of both, indicate that this model excels in achieving high accuracy across all classes. The CNN-RNN model demonstrated superior quality when compared to existing approaches. This implies that it has practical applications and can enhance methods of categorizing bottle gourds. This model’s accuracy surpasses the outdated manual methods of organizing things. This demonstrates the significance of improving agricultural labor and preserving diverse species. This study provides a precise method for categorizing different types of bottle gourds and also lays the foundation for utilizing combined CNN-RNN models in various agricultural scenarios, hence advancing crop sorting and farm technology.