This research delves into deep learning and machine vision applications for plant leaf disease detection in agricultural settings, focusing on farm village datasets. Utilizing a blend of authentic farm village data and synthetic data from Generative Adversarial Networks (GANs), three advanced convolutional neural network (CNN) models VGG16, ResNet50, and InceptionNet V3 are employed with transfer learning. Leveraging transfer learning enhances model performance through fine-tuning pre-trained networks. The study systematically evaluates models based on key metrics like accuracy, precision, recall, and F1 score. Results showcase the methodology's robustness, with ResNet50 emerging as the leading performer at 83.23%, contributing to precision agriculture's advancements with promising implications for sustainable farming and crop yield optimization.