Automated Pavement Crack Detection & Sealing System
The Georgia Tech Research Institute developed a prototype automated pavement crack detection and sealing system - called the Crackfiller - with funding from the Georgia Department of Transportation. In road tests, the system was able to detect cracks smaller than one-eighth-inch wide and efficiently fill cracks from a vehicle moving at a speed of three miles per hour.
Transcript
00:00:01 here at the Georgia Tech Research Institute we are developing a system to automatically fill cracks on the roadway currently crack seiling operations are an extremely manual process but the Georgia Tech system takes the humans entirely out of the loop with exception of the driver there is the safety issue with having several people walking around on the road during these crack
00:00:22 ceiling operations so the idea is that if you can automate that process you will help to make it safer plus more cost effective when they do crack seiling on the roads so the way the system works is first with the vision system what is unique about the vision system is we have a structured lighting approach so we have two sets of light emitting dodes or LEDs that are
00:00:45 directing light onto the surface of the roadway so they are doing this in an orthogonal fashion so that some LEDs are pointed in a different direction than the others by doing this we are able to bring out the cracks from the roadway and help ease the burden of processing on the image processing software so once we have identified the crack we now have to seal the crack and we are doing this
00:01:08 along the lines of how a dot matrix printer works so as we are driving along we then fill the cracks using the motion of the vehicle this only works for transverse cracks or ones that are perpendicular to the direction of the road we also have another part of the system which fills the cracks that are longitudinal or parallel to the direction of the road we have tested the
00:01:30 system system on numerous occasions once on a roadway in Georgia and at other times on parking lots and other surfaces that had a cracked asphalt we took thousands of images and by capturing several of these cracks we were then able to tune our image processing software in order to improve the efficiency of the crack detection software one of the major challenges of
00:01:53 the automated crack detection system and crack filling system is that it needs to operate continuously at mph the manual operation currently limits the way they do the traffic control because they have to stop the entire Lane because the process runs so slowly so that means everything is done in real time and that's what really sets this apart currently the crack ceiling
00:02:17 system works on a 1ft section of Road and we need to scale this up to where it fills a 14t wide section of road or a single Lane of pavement there are several challenges ahead but nothing that we f is unsolvable we feel that the current prototype system has demonstrated that it is feasible to fill cracks on a 14t wide section of Road it is possible to take an idea just a
00:02:41 simple idea and develop them fully into a system that will work and actually solve a problem that currently exists for the Departments of Transportation for