NAGOYA - Faint white road lines that fade over time are becoming a safety concern in Nagoya, prompting the city to test an AI-based system that automatically detects worn markings and helps determine where maintenance is needed most.
Drivers in Nagoya often see road markings that fade after about a year, posing a potential hazard. Until now, inspections relied on visual checks by city workers, but Nagoya has begun experimenting with a new approach using artificial intelligence. A smartphone is mounted on the windshield of a patrol car to record road surfaces. "The system automatically detects areas where lines have faded and uses that data to guide future repainting and analysis," said a patrol staff member.
The city, which manages roughly 6,600 kilometers of roads, has long depended on human inspection to monitor deterioration. However, officials admit that subjective judgment can vary between workers, leading to inconsistent results. "We oversee not only the pavement itself but also all related structures such as guardrails and street trees. Because inspections rely on human eyes, inconsistencies inevitably arise," said Makoto Kani, deputy chief of the Nagoya City Road Maintenance Division.
To improve accuracy, the city partnered with the Smart City Technology Research Institute, a University of Tokyo spinoff, to launch a pilot project last year. The AI app, installed on patrol vehicles, analyzes video footage and automatically evaluates the degree of fading on 15 types of road markings, including lane dividers, stop lines, and crosswalks, grading them on a four-level scale. "The yellow boxes indicate areas detected by AI as faded. The system evaluates every 10 meters," said company president Boyu Zhao.
The results are displayed on a digital map showing the degree of wear along each route, allowing maintenance teams to prioritize areas that require urgent repainting. "We’ve entered an era where everything visible to the human eye can be analyzed by AI. Areas marked in red indicate severe fading and can be addressed first," Zhao added.
The system, designed to suit Nagoya’s wide, multi-lane roads, can now detect fading across up to three lanes simultaneously. Although still in the testing phase across limited areas, expectations are high that AI will reduce workloads for city employees and standardize the quality of road markings.
The technology has already proven useful beyond Nagoya. After the Noto Peninsula earthquake in January last year, it was used on the Noto Satoyama Highway in Ishikawa Prefecture to assess cracked and damaged sections of road, helping authorities decide which areas to repair first. "Because repair budgets are limited, this system was extremely helpful in prioritizing where to allocate funds," Zhao said.
Nagoya plans to eventually equip all patrol operations with AI systems to enhance road maintenance efficiency and ensure the consistent upkeep of lane markings across the city.
Source: Nagoya TV News
