Massively parallel computing techniques might improve highway maintenance

Road maintenance procedures that allocate resources according to the condition of particular highway sections are more cost effective than those involving fixed-interval, budget-controlled maintenance. This more effective needs led approach requires measuring the highway's condition at regular, typically annual intervals. For this purpose, civil engineers build a highway management system into a database that records information about each section of roadway in the network under consideration. Traditional manual methods of measuring cracking in road pavements are slow and unreliable. In developing an automatic, repeatable and reproducible way of determining the extent of cracking, the authors have used a distributed array of processors to process digitized images at practical speeds.