The escalation of road infrastructure anomalies, such as speed breakers and potholes, presents a formidable challenge to vehicular safety, efficient traffic management, and road maintenance strategies worldwide. In addressing these pervasive issues, this paper proposes an advanced, integrated approach for the detection and classification of speed breakers and potholes. Utilizing a sophisticated blend of deep learning methodologies and enhanced image processing techniques, our solution leverages Object Detection to analyze and interpret real-time visual data captured through advanced vehicle-mounted camera systems. This research meticulously details the comprehensive process involved in the development of this system, including the acquisition and preprocessing of a vast, varied dataset representative of numerous road types, conditions, and environmental factors. Through rigorous training, testing, and validation phases, the model demonstrates remarkable proficiency in recognizing and classifying road imperfections, outperforming existing models in accuracy, precision, recall, and F1 score metrics.The innovative approach underscores the potential of integrating cutting-edge machine learning algorithms with real-world applications to significantly enhance autonomous navigation systems and assist in the proactive maintenance of road infrastructure. Furthermore, the paper discusses the model's scalability and adaptability, emphasizing its potential application across a wide range of environments, weather conditions, and vehicular technologies. By providing a detailed analysis of the system's performance under varied operational scenarios, this research contributes valuable insights into the realm of intelligent transportation systems and smart city infrastructure development ,the findings of this study not only pave the way for future advancements in autonomous vehicle technologies but also offer promising perspectives for the improvement of road maintenance and safety protocols. Looking ahead, the paper outlines a series of potential enhancements and areas for future research, including the optimization of the detection algorithm for increased efficiency and the exploration of its integration with other sensory data sources for a more holistic approach to road anomaly detection.