Autonomous vehicles rarely falter because their sensors can’t perceive; it fails when the road it sees is unreadable. This talk presents a lightweight camera-plus-INS pipeline that runs on production vehicles, flags faded lane paint, damaged cable barriers, and missing or obscured traffic signs, and distills those detections into practical KPIs. The results flow directly into an OEM’s internal map, letting perception stacks continue to recognize and trust roadway assets even when they degrade. Engineering dashboards reveal emerging trouble spots early, giving ADAS teams time to refine algorithms or update maps before issues turn into field faults. Because the system relies on sensors already present in most fleets, it scales quickly, and the same data can be shared with DOTs to accelerate repairs and improve overall road safety.