Automatic Detection of Pavement Hazards Using AI
Wide field of Feature
ㆍHigh-resolution imaging at high speed
Captures high-resolution images without blurring or distortion, even at driving speeds of up to 100 km/h.
ㆍHigh precision and recall performance
Conducted in-house performance testing over a total driving distance of 716.8 km across sections managed by the Korea Expressway Corporation. Test results achieved 90% precision and 99% recall.
ㆍReduced road accident risk and maintenance costs
Enables automation of road damage detection tasks. Eliminates the need for road closures, reducing traffic congestion and accident risk. Establishes a smart pavement maintenance system, leading to reduced road repair budgets.
Various Fields
ㆍAutomatic detection of pavement hazard factors on asphalt national highways and expressways
ㆍAutomatic detection of pavement hazard factors on concrete national highways and expressways
ㆍReal-time detection of pavement damage and road surface image acquisition
ㆍGeneration of pavement condition maps based on pavement image data
ㆍ"Pavement damage" refers to the deterioration or defects on roads or paved surfaces, which significantly affects the operational stability of high-speed vehicles
ㆍRoadVision utilizes pre-trained AI models to perform real-time detection of pavement damage, including pavement crack, surface defects, and damage to repair materials.
ㆍThe detailed detection types of detectable pavement damage are as follows.
ㆍPavement condition map is an image-based map used to visually assess the condition of roads immediately by stitching vertically captured road images together, identifying areas in need of maintenance.
ㆍIt utilizes colors or symbols to represent the extent of road damage, displaying the size of road damage differently based on depth or size.
ㆍHigh-resolution vertical road images are generated and uploaded in real-time to a server using 5G communication to automatically create the pavement condition map.
Korean
Russian