Leaf Scanning Disease Detection

Crop diseases pose a significant threat to global food security, with rapid identification challenges in regions lacking essential infrastructure. The confluence of rising smartphone adoption worldwide and recent strides in computer vision through deep learning has opened avenues for smartphone-enabled disease diagnosis. Leveraging a public dataset containing 2,500 images of plant leaves in varied health conditions, acquired under controlled settings, we employed a deep convolutional neural network. This model successfully discerns crop species and identifies diseases or their absence, achieving an impressive 86.35% accuracy on a withheld test set. This underscores the viability of the proposed methodology. In essence, the strategy of training deep learning models on expansive and accessible image datasets signals a promising route for widespread smartphone-assisted crop disease diagnosis on a global scale. Keywords - crop diseases, machine learning, deep learning, convolutional neural network (CNN)