Recognition of the need for good pavement data has resulted in efforts to develop noncontact automated pavement data acquisition systems. These systems collect roughness, geometric, and distress data using a diverse range of sensing technologies. Although many of these technologies are becoming well developed and accepted, there has been a lag in the development of effective methods to exploit the data that are produced. To do this, the diverse technologies used for pavement sensing must be integrated. Currently, integration methods are typically ad hoc and specialised for each application and sensor type. A general method for such integration is described. This general method supports diverse sensors, large datasets, and data abstraction over a wide range, from the location of cracks of a fraction of an inch in width to summary pavement section condition measures. Also, integration requirements are examined, issues of uncertainty are discussed, and examples are presented of pavement sensor data combinations for two diverse application areas. One example integrates video and laser range data to identify pavement cracks. The second example illustrates sensor data integration for pavement condition assessment using three commercial survey systems. Data from these three commercial systems are integrated using the same software model structure and typical summary statistics are derived.
Samenvatting