This paper provides a complete solution to the most crucial road safety enhancement by proposing real-time object detection, lane detection, and driver alerts in single framework. The framework uses deep learning methods to jointly identify and segment different road objects (vehicles, pedestrians, animals, stones) in complex road scenarios. The system achieves the real-time identification of these hazards by combining YOLOv5, YOLOv8, Mask R-CNN, and the CNN-based lane detection module to accurately detect several types of potential hazards and supply timely alerts to drivers. The lane detection module provides high accuracy in collecting lane markings based on advanced image processing and robust CNN for diverse conditions. The proposed methodology, which is better in comparison to existing methods, can lead to higher-level functions like advanced driver assistance systems (ADAS) and autonomous cars.