An automated pavement crack detection and sealing system equipped with LEDs and a camera increases worker safety, requires less manpower, and increases the amount of roadway covered per day. The prototype system developed at Georgia Tech Research Institute (GTRI) can efficiently detect and fill cracks smaller than 1/8-in. wide from a vehicle moving at a speed of 3 mph. An automated crack sealing system would increase the level of safety for the workers involved, require fewer personnel, and increase the amount of roadway covered per day. (Images: Jonathan Holmes) Mounted on a trailer, the automated system consists of a stereo camera, LEDs in two different colors, and an assembly to provide a continuous supply of sealant to longitudinal and transverse sealant distribution systems. The operation requires only one worker, who drives the vehicle pulling the trailer. As the system travels down a lane, LEDs illuminate the road in two directions — parallel and perpendicular to the road — while the stereo camera takes two pictures of the road simultaneously, which are analyzed using thresholding and filtering algorithms. Less than 100 ms after the pictures are taken, a computer onboard the trailer generates a “crack map” specifying the location and shape of any cracks shown in the images. “Our prototype system has proved in many ways that a commercial-scale automated crack sealing system is viable,” said Jonathan Holmes, the GTRI research engineer currently leading the project, which began in 2003. “We demonstrated solutions to technical challenges — including the high-speed firing of nozzles, automated crack detection and navigation — in a real-time, limited-scale system.” Example of a pavement crack filled with GTRI’s prototype automated pavement crack detection and sealing system. Based on the information on cracks generated from the map, the master controller can instruct the sealant applicator valves when to fire. For longitudinal cracks, a single dispensing nozzle capable of continuous operation was attached to a linear servo axis. The transverse sealant distribution system consists of 12 nozzles spaced evenly across one foot. The transverse and diagonal distribution prototype represents a module that can be replicated and joined together to service a full-width roadway lane. The system proved successful for automatic crack sealing operations in multiple road tests. Before a full-scale system can be implemented by transportation departments, the scientists plan to improve the crack detection algorithm. The current algorithm tested more than 100,000 images, correctly identifying more than 83 percent of the cracks. In road tests, the system detected cracks smaller than 1/8-in. wide and efficiently filled cracks from a vehicle moving at a speed of 3 mph. “Our crack detection algorithm was limited because we used a vision-based system, which was confounded by regions of high contrast caused by features other than pavement cracks, including dark stains in the pavement, lane stripes, raised-pavement markers, crack sealants and debris,” Holmes said. “A full-scale system may require a fusion of multiple imaging sensors, such as a 3-D laser scanning system.” The team will also look at how the sealant was supplied to the longitudinal and transverse distribution systems before a full-scale system is rolled out. The prototype was funded by the Georgia Department of Transportation. For more information, visit: www.gtri.gatech.edu