Leveraging Convolutional Neural Network for Classification of Leaf Disease

The population in the world has increased and the need to improve agricultural productivity has become several times greater, presenting significant challenges to farmers. Crop diseases, especially those occurring in the leaves, remain one of the most significant risks to crop health and production. Crop diseases can drastically reduce agricultural production, putting food security at risk everywhere. This study addresses this challenge by capitalizing on the potential of advances in machine learning, and especially that offered by capabilities of Convolutional Neural Networks (CNNs), and getting combined with the processes involved with image processing techniques. A minimum error monitoring set up done periodically, build an automatic plant disease-identifying system relying entirely on an extensive database containing images of labeled leaves together with respective diseases. Some major crop diseases are effectively classified using CNNs, because of the strength in recognizing patterned structures and features within an image. Methods such as transfer learning and augmentation of data are some of the methods we will implement, so the models are able to produce diagnoses under various conditions. In other words, the ultimate goal of the project is to offer a strong earlydetection solution to plant diseases. This will enable farmers to take preventive measures and minimize potential losses by receiving timely and accurate information on crop conditions. This innovation will help in promoting sustainable farming practices and contribute towards improved food security worldwide. The model demonstrates strong performance in identifying the healthy leaves (precision <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$=0.93$</tex>,recall <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$=0.92$</tex> and F1 score = 0.90) and with moderate success in identifying powdery leaves (precision <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$=0.85$</tex>, recall <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$=0.91$</tex> and F1 score <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$=0.87$</tex>) and rust leaves (precision <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$=0.89$</tex>, recall <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$=0.84$</tex> and F1 score <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$=0.86)$</tex>.