Ensuring the dependability and safety of autonomous vehicles in real-world driving conditions depends heavily on accurate lane detection. Adverse weather conditions, poor illumination, and other challenging environments can significantly compromise the performance of lane detection systems. This study presents a deep learning based data augmentation approach to improve the robustness of lane detection models under such conditions. Synthetic training data is generated by simulating adverse scenarios such as fog, rain, and glare on real driving images. These variations help the model generalize better across different environmental situations. The augmented images are used to train a convolutional neural network, and the model is evaluated using a benchmark dataset of actual driving scenes. Performance is measured using accuracy, precision, recall, and F1 score. The results show that applying data augmentation during training significantly enhances lane detection performance and improves the overall reliability of autonomous vehicle navigation in adverse weather conditions.