By KHT Staff. AI-generated illustration of pothole detection, based on a city demonstration image published by iThome.
KAOHSIUNG — Cameras on buses and other vehicles are helping Kaohsiung find potholes, with one reported case taking about two minutes from detection to assignment of a repair job.
The city presented the results on Oct. 29 as part of its AI Lighthouse project (高雄AI燈塔計畫), which uses artificial intelligence to flag traffic problems, damaged roads, flooding, and environmental hazards.
According to Chunghwa Telecom (中華電信; TWSE: 2412), a project partner, 45 vehicles equipped with 135 cameras collect road images. AI identifies possible defects and sends their locations to the responsible agencies. The two-minute example measured the time to assign work, rather than complete a repair.
For traffic management, the company said the system issued alerts about unusual congestion 14 minutes earlier than conventional manual recording. It combines existing traffic systems with AI image analysis to help staff identify incidents and possible causes.
Flood monitoring draws on more than 400 monitoring stations and access to imagery from more than 13,000 cameras. The city government said AI proactively detected 52% of the events recorded during heavy rain in August.
Environmental applications flag smoke, open burning, suspected illegal dumping, and roadside rubbish. Chunghwa Telecom said alerts containing images and locations could be sent within 30 seconds of detecting an abnormality, helping staff investigate and arrange a response.
As Formosa English News reported, the shared platform lets departments follow an incident’s location, footage, and handling progress, connecting camera detections with the people responsible for responding.
