Sustainable Farming: Automating Apple Disease Detection with Max-Voting Ensemble Deep Learning

The Food and Agriculture Organization of the United Nations (FAO) estimates that 20 to 40 percent of crops produced around the world are lost to pests, costing $220 billion dollars annually. With a high nutritional and medicinal value, apples are one of the key fruits for healthy living today. Most industrialized apple orchards currently rely on human vision for disease detection, which causes a significant lag in the tracking of apple diseases resulting in poor crop yields and fruit quality. Automating apple disease detection using machine learning will lead to sustainable farming. Deep learning, a branch of machine learning, is well-suited for learning from image data. In this paper, Convolutional Neural Networks (CNN), a type of deep learning model, are investigated to accurately classify healthy and disease-affected apple leaf images. A baseline CNN model was developed along with other CNN models, namely DenseNet121, EfficientNetB7, MobileNetV2, ResNet50, and VGG16. A max-voting ensemble that included the most accurate CNN models was then deployed on the web. Consequently, apple farmers and other users can detect the three common apple leaf diseases – apple scab, black rot, and cedar apple rust, as well as healthy leaves by uploading their own images from the orchard.