Perception-Oriented Visibility Assessment of Road Signs Using 3D Point Clouds

Road signs are critical components of road infrastructure that support safe and efficient traffic operations. With the rapid advancement of intelligent transportation and automated driving technologies, increasingly stringent requirements are being placed on the visibility and effective presentation of road-sign information under diverse traffic environments. However, existing visibility evaluation methods mainly focus on geometric visibility or object detection accuracy, while the effective perception of sign information under complex roadside environments remains insufficiently addressed. This study proposes a perception-oriented visibility assessment framework for road signs using three-dimensional point clouds. A LiDAR-based three-dimensional road model is first established, and traffic signs are automatically detected from synchronized video images using YOLOv10s. To differentiate the predefined color-based importance of different sign regions, a weighting strategy based on traffic sign color composition, typical functional roles, and visual-cognition findings is introduced into the visibility assessment process. The proposed framework extends conventional equal-weight geometric occlusion assessment by quantifying the contributions of different sign regions under complex occlusion conditions. A highway reconstruction and expansion project is employed as a case study to demonstrate the application of the proposed framework. Across the four evaluated signs, the maximum occlusion rates ranged from 0.25% to at least 30%, with three signs classified as Grade 3 or Grade 4 occlusion. The corresponding image-based detection distances ranged from 30 to 90 m, while field images qualitatively confirmed that vegetation was the primary obstruction affecting the three occluded signs. The proposed method provides quantitative visibility information for road sign inspection and maintenance, helping identify visibility deficiencies and supporting more targeted maintenance interventions. By facilitating the timely management of roadside obstructions and reducing reliance on purely experience-based inspection, the framework can contribute to safer, more efficient, and sustainable management of road infrastructure.