Innovative Roadguard: Yolov8-Powered Road Damage Detection

Road damage poses a significant threat to public safety and infrastructure integrity. The proposed project enhances road maintenance by employing YOLOv8, a state-of- the-art deep learning model, for accurate and efficient identification of various types of road damage. Existing methods for road damage detection predominantly rely on traditional computer vision approaches or earlier versions of YOLO that often struggle with real-world challenges, such as variations in lighting conditions, diverse road surfaces, and the presence of complex patterns. The innovation in this project involves the integration of advanced pre-processing techniques, data augmentation strategies, and fine-tuning procedures specific to road damage scenarios. Its single pass detection provides speed and accuracy to enhance the model's real time detection capabilities. The proposed method enhances accuracy for precise road damage identification, aiding effective resource allocation. Sufficiently high mAP values indicate the robustness of the model under varied environmental conditions. There is also a comparison done using the yolov5 model on the same dataset and exact some conditions to prove the efficiency of yolov8 in practical situations.