Detecting Crop Pests and Diseases Through Deep Learning Techniques for Improved Yields
摘要
To optimize agricultural operations and guarantee food security, crop categorization relies heavily on the precise identification of a wide range of agricultural goods. This study offers a thorough assessment of a crop classification model, concentrating on the classification performance of four different crop categories: cashew, maize, tomato, and cassava. While the model successfully classifies cashew (79.89%), maize (61.60%), and tomato (43.28%) with outstanding accuracy, it has substantial difficulties with cashew. With a classification accuracy of barely 14% and unusually high loss values, cassava is the most significant challenge. These findings underline the difficulty of correctly differentiating certain crops, notably cassava, and emphasize the need to improve and fine-tune the model. Future research endeavors should prioritize improving the model’s accuracy and generalization, aiming to create a robust and reliable crop classification system that can effectively support agricultural communities and enhance food security.