They say an ounce of prevention is worth a pound of cure, but when it comes to road maintenance, an ounce of prevention is worth several tons of tarmac. A tiny crack in the asphalt may not seem like much, but once it lets in rain and frost, it’s a ticket to potholes and a very expensive resurfacing. The problem is that crack repair is time consuming and labor intensive, so the Georgia Tech Research Institute (GTRI) has come up with an automatic pavement crack detection and repair system that operates at comparable speeds to conventional methods, but with fewer people and less exposure to hazardous fumes.

Road resurfacing, which involves tearing up old tarmac, repairing roadbeds and laying down a new surface is expensive, time-consuming, labor intensive and can be a massive disruption for motorists. Because of this, regular maintenance to deal with roadway cracks is vital to keep traffic flowing smoothly and roadwork costs down. Unfortunately, conventional methods for dealing with cracks also require a relatively large crew and the work is boring, repetitive and exposes workers to hazardous fumes. For this reason, a Georgia Tech team led by GTRI research engineer Jonathan Holmes are developing an automated system that identifies, maps and repairs road cracks on its own.


More than 1,200 New Atlas Plus subscribers directly support our journalism, and get access to our premium ad-free site and email newsletter. Join them for just US$19 a year.


Currently, the scope of the project is limited to dealing with thin cracks that require only simple repairs rather than the larger ones that indicate roadbed damage and need more extensive work. This approach makes the development task manageable because the repair of thin cracks only requires the application of a liquid sealing compound.

LED imaging system

Unlike conventional methods, the Georgia Tech system needs only one operator. Since the only thing the operator does is guide the machine down the road, this is not only a major savings, but the operator is also safe from toxic fumes given off by the sealant.

The machine manages this by means of a stereoscopic imaging system that automatically identifies and maps the cracks. The system uses two sets of LEDs aligned parallel and perpendicular to the road surface. Each LED set is in a different color and a stereo camera takes two images. Based on the contrasting information produced by these two colors, a thresholding and filtering algorithm generates within 100 milliseconds a map of any cracks as small as an eighth of an inch (3 mm) in width. This map is used by the machine to guide an array of twelve nozzles that efficiently spray sealant into the cracks. The whole system can operate at a speed of up to 3 mph (4.8 km/h)

The Georgia Tech system is still in the proof of concept stage of development. “Our prototype system has proved in many ways that a commercial-scale automated crack sealing system is viable, said Holmes. "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.” The imaging system is still in need of improvement with only an 83 percent success rate for mapping and the imaging algorithm is in need of tweaking. Holmes also believes that the sealant supply system could be improved.

Currently, the machine can only operate on a narrow strip of highway, but it was designed to be modular, so building a larger version that can handle an entire lane or road at a time is a simple matter of scaling up. If this is possible, then the device could have great potential if it can be extended to handle more aspects of crack mending. Repairing cracks, especially larger ones that may contain debris and dealing with complicated “alligator” cracking is more than just squirting sealant. If the Georgia Tech system can be built deal with general crack repair, then at least one road maintenance job will be in future reduced to the same level of motorist aggravation as the passing of a street sweeper.

Source: GTRI

View gallery - 3 images