This paper introduces an innovative methodology utilizing image analysis with neural networks (NN) to enhance driver awareness and mitigate road safety risks. By employing Scenic and CARLA simulators, we generate diverse synthetic images representing various driving scenarios, ensuring robustness and generalizability. Unlike traditional approaches, our parameter-free NN model autonomously learns to classify safe and unsafe distances from input images without manual calibration. Through deep learning, it discerns patterns effectively, offering real-time alerts for potential hazards. This methodology enables proactive intervention precisely when driver attention is critical, fostering a symbiotic relationship between driver and automation. We present three datasets comprising 12,000 synthetic images each, establishing the first synthetic dataset for distance classification. Experimentation with custom CNN, VGG16, and ResNet-18 demonstrate the efficacy of our approach, with specified distance ranges for safe and unsafe scenes. Our system shows promising results in classifying distances, contributing to improved road safety.