Pedestrian safety remains a critical concern in urban environments, with vulnerable road users accounting for a significant portion of traffic fatalities worldwide. This study presents the design, implementation, and evaluation of a human-centered in-vehicle visual interface integrated into a Pedestrian Protection Advanced Driver-Assistance System (ADAS). Leveraging the CARLA simulator, the system incorporates sensor fusion between RGB and depth cameras, alongside a YOLOv8-based detection pipeline, to enhance real-time pedestrian recognition. A participatory design process, including mockups and focus group sessions, guided the development of the interface, which uses multimodal alerts and intuitive visual cues to convey risk levels and improve driver situational awareness. The interface divides the driving scene into safe, warning, and danger zones, dynamically updating alert information based on estimated time-to-collision. Evaluation through user testing and the User Experience Questionnaire (UEQ) demonstrated strong usability, particularly in terms of perspicuity and efficiency, confirming the system’s effectiveness in supporting timely and informed driver responses without causing cognitive overload.