The optimal route between destinations at a given time depends on the traffic conditions and speed limits at that time. Detailed and accurate data on traffic conditions and speed limits is sometimes unavailable for a subset of road segments within potential routes. This disclosure describes techniques to determine various parameters about the likely prevailing travel conditions for a given road segment at a given time based on satellite images. The inferred travel conditions can be appropriately factored in when suggesting routing guidance. The satellite images are processed via appropriately trained machine learning models that can detect relevant attributes or physical features. An ensemble model composed of individual trained machine learning models is utilized to analyze the detected attributes and physical features to provide an estimate of road conditions likely to be present at travel time.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.