EARLY DETECTION OF FUNGAL DISEASES IN CROPS

The early detection of fungal diseases in crops is crucial for timely intervention and effective crop management, reducing yield losses and promoting sustainable agriculture. This study presents a deep learning-based system for the early identification of fungal infections in crop leaves, utilizing a structured process to improve accuracy and efficiency. A leaf dataset is compiled, consisting of images of both healthy and infected leaves. Each input image undergoes a preprocessing stage that includes resizing, noise removal, and contrast enhancement to improve the quality of data for analysis. A Convolutional Neural Network (CNN) is then employed for feature extraction, learning patterns that distinguish healthy leaves from those infected by fungal diseases. Following feature extraction, the CNN classifies the leaves into two categories: healthy or infected. In addition to visual classification, the system is equipped with a voice output feature that announces the diagnostic result, making it user-friendly for farmers and agricultural workers. The model achieves a high accuracy rate in detecting fungal infections, showcasing the potential of deep learning in precision agriculture. This approach provides a fast, reliable, and non-invasive method to monitor crop health, enabling earlier treatment and reducing the need for harmful chemicals.

Keywords: Leaf Dataset, Deep Learning, Convolution Neural Network, Image Processing Techniques and accuracy.