Detecting road cracks is vital for maintenance and safety, as they pose risks like pavement degradation and structural instability, endangering motorists. Timely identification through our novel method, integrating image processing and machine learning, aids prompt intervention, averting further deterioration and potential accidents. This study outlines a comprehensive process commencing with drone-captured road surface imaging, followed by preprocessing steps like Gaussian filtering for noise reduction and contrast enhancement to enhance crack visibility. Subsequently, the pre-processed images undergo feature extraction, focusing on edges, textures, and specific crack-associated patterns. These extracted features are then inputted into a machine-learning model trained on a diverse dataset of road images, facilitating robust crack detection. The YOLO (You Only Look Once) v8 model is also applied to detect road cracks, a critical infrastructure maintenance and safety task. This model uses deep learning techniques to swiftly and accurately identify cracks in road surfaces. By employing YOLO v8, the detection process becomes more efficient, enabling rapid assessments of road conditions. This model’s performance surpasses traditional methods, particularly on Indian road surfaces, significantly improving accuracy in identifying and locating cracks. This innovative approach fosters effective road infrastructure management, ensuring the safety and durability of transportation networks.