Deep Convolutional Neural Network-based Framework for Apple Leaves Disease Detection

Apples are a popular fruit all around the world, nevertheless, they are mostly farmed in Asia. Moreover, approximately 76 million tons of apples are produced annually around the world. Furthermore, Apples may aid in preventing cancer, metabolic syndrome, cardiovascular disease, diabetes, and a variety of other diseases. However, various environmental conditions and other factors affect Apple leaf plant growth. In addition, the primary cause of the production crisis is apple plant disease. ATLDs such as RUST and SCAB are all popular and significantly impact apple leave yield. Therefore, numerous researches have been carried out to detect apple leave diseases automatically. However, there is still room for improvement in efficiency, computation complexity, time consumption, cost, and variety of techniques. This research employs deep convolutional neural network models VGG-19, ResNet-34, and Dense-121-Net to identify apple leave diseases. Besides, pre-processing of the dataset images enhanced the image quality and remove the noise. Furthermore, Data augmentation methods are also utilized to expand the number of images in the dataset. Moreover, models employing VGG-19, Resnet-34, and Dense-121Net are analyzed through the plant village dataset and attained 98.02%, 97.06%, and 99.75% accuracy respectively. An evaluation of networks in the plant village dataset shows that the developed algorithm performs better and has an advanced methodology suitable for real-time agricultural disease detection applications.