Assessing the Effectiveness of Augmentation Techniques in Enhancing Plant Leaf Disease Classification

Plant leaf disease identification is vital for food security, as plant diseases cause significant agricultural losses. Early detection of symptoms on plant leaves is crucial for minimizing yield loss. Traditional monitoring is labor-intensive, prompting the use of deep learning for automated detection. However, the lack of large-scale, high-quality, open datasets remains a challenge, with many being closed-source or suboptimal. Image augmentation techniques can expand dataset size, improving model performance without the need for additional data collection. In this work, we explore the impact of various augmentation techniques on the performance of deep learning models, particularly in lab and field datasets. Our results show that techniques like color, transformation, and noise augmentation significantly enhance model accuracy, with combined augmentation yielding the highest accuracy, especially for field datasets. These findings underscore the effectiveness of augmentation in improving deep learning models for plant leaf disease identification.