In This Space
Transport Network & Road Status
Transport network and road status are essential datasets for logistics planning. As narrow linear features, the detection and extraction of roads and their attributes is a challenging task. The base imagery required is typically high to very high resolution optical data since such sensors provide the spatial resolution required to detect road features.
Baseline transport network information includes roads and road types (single track, multi track), railways or railroads including track gauge, ferries as well as airports.
Baseline road status information includes the road surface (e.g. sealed or un-sealed, gravel, dirt) and road size (single-track or multi-track). Other information could include the road elevation gradient (for which elevation data are needed) and/or nature of turns along the road.
Information on transport network and road condition is important during the wet season in tropical countries, where seasonal rains can prevent travel along certain routes. In mountainous regions, roads may regularly become impassable due to landslides, especially in areas where roads are former logging roads. River crossings may also be affected. The degree of deterioration of a road surface (pot holes, rutting, etc.) is difficult to detect with current spatial resolution of satellite sensors.
Road condition and status is typically assessed using very high to high resolution optical data.
Known restrictions / limitations
Limitations for use of optical and radar sensors include the extraction of road baseline information and status in forested areas where the canopy can prevent visibility of the road surface. In mountainous areas, radar sensors are not likely to perform well because of sensor geometry.
Manual extraction of baseline road geometries may be required to provide accurate results, but is more time consuming and expensive than automatic methods. Management of conflation issues can be time-consuming where multiple data sources result in conflicting road information.
Monitoring roads requires frequent data acquisitions at high spatial resolution. Along with the linear nature of roads, this can result in high acquisition costs. Cloud cover also affects optical datasets.
Lifecycle stage and demand
Pre-licensing and Exploration:
Development and Production:
Geographic coverage and demand
Demand is global, focusing on less developed countries or remote areas. Information is most desirable on the status of mountain roads and roads within forested areas that have unpaved/dirt surfaces. Floodplains with a history of frequent flooding would be a focus for seasonal monitoring.
OTM:029 Pre-licensing site selection
Input data sources
Optical: VHR1,VHR2, HR 1
Radar: VHR1, VHR2, HR1
In particular, new sensors such as SkySat-1/2 PAN and Planet Labs may provide the required combination of spatial resolution and temporal frequency.
Obligatory supporting data:
Spatial resolution and coverage
Spatial resolution: 0.5 m – 10 m pixel size
Detailed transport network and road status detection and classification require VHR1 to HR 1 data.
Major roads can be detected using HR2 images such as Landsat 8.
Minimum Mapping Unit (MMU)
The MMU is based on the expected minimum size of the road features to detect. Four metres is a typical single-track road width, although the contribution of road reflectance in coarser imagery usually allows for estimation of the road centreline.
The minimum feature width (MFW) could be as fine as a single-track road with a width less than two metres.
Accuracy / constraints
Thematic accuracy: Accuracy of road detection should be >90% in areas of low vegetation cover and density. Accuracy can degrade based on topographic effects and vegetation cover. Expected accuracy is to be defined for specific projects. The accuracy of road condition and type attributes is expected to be >75% depending on specific project areas.
Spatial accuracy: The goal would be 1 pixel, but this depends on reference data.
Accuracy assessment approach & quality control measures
Automatic transport network and road status detection and classification should be checked against ground validation data (e.g. GPS data from vehicle or foot travel) and visual interpretation of road surfaces. Ancillary vector data from local or regional government or industrial operations (forestry etc.) are sometimes available.
Stratified random points sampling approach can be performed.
Frequency / timeliness
The frequency is constrained by satellite revisit and acquisition timeframes, but also processing requirements. While the minimum frequency is technically driven by the revisit cycle of the satellite, the maximum frequency is defined be the customer. Depending on the requirements of the customer the most suitable satellite sensor has to be selected, considering spatial / spectral resolution as well as revisit frequency. Most of the time, long-term changes are detected in 2 years or longer intervals (frequency can be lower depending on demand).
Baseline road mapping is a one-time process. Monitoring of changes in road status is typically performed on an annual basis. Daily change detection may be required during seasonal flooding or in emergency situations, if possible.
Timeliness of delivery:
Baseline data will depend on the availability of recently archived imagery or new image tasking. Processing is not challenging and products can be available in 1 week to 1 month depending on project size. Manual road extraction can be more time consuming and depends of many factors (area, designed accuracy and level of detail). Change detection and updates to validated road networks can be performed quickly.
Emergency assessment of road conditions can be completed in near real time (< 24 hours) depending on the established processing chain for the project and availability of base images.
On-demand availability from commercial suppliers.
New acquisitions can be requested globally.
Archived products availability may be limited for specific dates and locations.
Delivery / output format
|Peer Reviewer:||Hatfield Consultants|
Andy Dean, Sebastian Aleksandrowicz, Maria Lemper
Transport Network & Road status
# of Pages:
Internal – Project consortium and science partners
External – ESA
This page has no comments.