Road surface analysis

Automatic detection of road defects

Imagine that organizations no longer have to send inspectors on the road to fix road defects, and  instead damage to roads can be detected with artificial intelligence. And that the severity, extent and classification of the defects can be determined in an objective way, based on high resolution imagery. This has a major impact on the efficiency of road inspection.

Arcadis and Cyclomedia

Unique collaborations result in unique products. CycloMedia and Arcadis have combined their knowledge and automated the work of a road inspector through artificial intelligence. CycloMedia is responsible for the annual capturing of high resolution panoramic imagery and LiDAR point clouds. As a leading an expert in road analysis, Arcadis developed an algorithm that allows road defects to be automatically detected.

Size and position

The LiDAR point cloud of Cyclomedia is an very detailed, 3D representation of public space. With this photorealistic data source, the size and position of a road-defect can be determined exactly.

Budget efficiently spent in the right place

With the algorithm, road segments of 100 ft and 1000 ft can be selected and classified according to the guidelines of ASTM and PASER. This provides direct insight into which road segments should be maintained first. The available budget for road improvement will be spent where it has the biggest impact for road users.

Experienced road inspectors

The algorithm is trained and validated by Arcadis' experienced road inspectors. These road inspectors contributed their in-depth knowledge of the common methods for classifying road defects inserted into the algorithm. This ensures a constant quality of detection of road defects.

Uniform, objective and constant

An algorithm doesn't know good days or bad days. The results of the analysis are always objective and completely uniform. Differences in interpretation of road defects do no longer occur. In addition, the analysis can be repeated annually, making a year to year comparison possible.