Road Surface Inspection

Fully automated road inspection solution performs detection of surface defects based on high-speed optical acquisition systems with the following accurate defects classification.

To acquire actual detailed information about surface conditions of thousands of kilometers of state roads, the whole process needs to be fully automated. BitRefine group offers a solution for high-speed processing images of road surfaces into structured geolocated data, where manual defects labeling is no longer needed. Based on a trained convolutional network, our model belongs to the third generation of defect classification systems. It is capable not only of detecting cracks, but also of performing full classification of road damage the way a human would.

The most important features of our automated road inspection solution are:

• High accuracy in different conditions
• Full classification system recognizing objects such as patches
• Precise output data
• High-speed parallel computing with support of GPU
• Offline as well as online processing
• Integration with existing databases

Our recognition core is capable of processing images, captured by different cameras, installed in different ways. This can be installation on a special road inspection vehicle, as well ordinary cameras installed on public transport vehicles such as buses. Such options can also be combined into one system, which takes advantage of having precise data with measurements from special vehicle and constantly updating rough data from public transport. BitRefine provides reporting infrastructure as well as services of integration into an existing system.