Apple Leaf Disease Detection using DenseNet121 Transfer Learning Model

Abstract

Apple tree cultivation is important in global horticulture because of its contribution to food production and economic benefits. Nonetheless, leaf diseases such as discoloration, lesions, and deformities endanger the vitality and efficiency of these trees. The prompt emphasises the importance of identifying these diseases quickly and precisely in order to effectively implement control measures that are specifically tailored to protect food production quality. There has been a notable convergence between advanced technologies and the agricultural sector in recent times, resulting in a transformative impact on disease management. In this context, deep learning models have emerged as an effective tool. The main goal of this research is to classify apple leaf diseases using a fine-tuned DenseNet121 Transfer Learning Model. The model distinguishes between four primary disease categories, namely "Healthy," "Multiple Disease," "Rust," and "Scab." Because apple leaf diseases are diverse and intricate, categorising them into meaningful groups is critical. Different disease groups exhibit distinct symptoms, necessitating the expertise of a skilled observer and a thorough understanding for accurate differentiation. Traditional diagnostic techniques that rely heavily on human observation and visual cues frequently yield inconsistent and imprecise results. The emergence of deep learning, a subset of artificial intelligence, has ushered in a new epoch in the field of disease identification. The model was trained for a total of 26 epochs, with each epoch consisting of 32 batches. During the training process, the Adam optimizer was used. The outcome is evaluated by taking precision, recall, F1 score, and accuracy into account. The model's classification accuracy is exceptional, reaching an impressive 99%. The primary goal of this article is to add to the growing body of knowledge in the field, thereby encouraging further research and innovation in agricultural technology.