One of the goals was to filter small single cracks not affecting significantly road quality and only pay attention to bigger damage areas.
Road maintenance is a complex task requiring:
- Localization of defect areas
- Counting approximate area sizes for repair
- Planning sufficient but not excessive material and human resources
- Mission Planning with UgCS
- Drone: DJI Phantom 4
- Map storage and processing: ATLAS
- Initial machine learning detector's training in ATLAS cloud: approx. 50 minutes
- Shareable web-map with marked defects in ATLAS
- Report with georeferenced areas in GeoJSON
Here is an example of how areas with cracked asphalt can be identified with Atlas: CASE#1
Client: G2 INGENIERIA PARA LA CONSTRUCCIÓN S.A. DE C.V.
Industry: Civil Engineering, Construction & Maintenance