The olive tree is susceptible to various diseases that affect its health and yield, making early and accurate detection crucial for effective management. Deep learning has emerged as a powerful tool for automating disease detection, but its performance depends on large, balanced datasets. A key challenge is the scarcity of images for certain diseases, leading to class imbalance and potential bias in model predictions. This study addresses this issue by using a data generation technique involving (1) classical image transformations and (2) Deep Convolutional Generative Adversarial Networks (DCGAN). Experimental results using a pre-trained Convolutional Neural Network (CNN) for classifying olive tree diseases demonstrate that data balancing significantly enhances classification accuracy, highlighting the importance of data generation in deep learning-based diagnosis. This approach improves automated disease detection and can be extended to other plant pathology applications.