9.8.3 — Urban Mobility Systems — maturity: live
Curbside Activity Mapping
Using very-high-resolution satellite imagery and machine learning to classify and quantify curbside occupancy — parked vehicles, loading zones, ride-hail staging, and illegal stopping — across an entire city.
Satellite imagery and AIS-derived analytics let city governments count loading bays, track ride-hail dwell times, and enforce kerb-use policy without installing a single roadside sensor.
City transport planners are largely blind to what happens at the curb. Loading bays are blocked by private cars, ride-hail vehicles cluster illegally, and bus stops are obstructed during peak hours — all generating congestion, emissions, and safety incidents that ground-level enforcement cannot monitor at scale. Traditional methods (manual counts, fixed cameras) are expensive, patchy, and politically contentious when contracted out to foreign vendors who own the data.
A constellation of very-high-resolution (sub-0.5 m) optical microsatellites provides systematic, city-wide snapshots of every arterial street multiple times per day. Object detection models classify vehicle types — heavy goods vehicles, vans, passenger cars, buses, motorcycles — by kerb position, dwell time proxy, and lane-of-standing. Revisit frequency is too coarse for real-time enforcement but is ideal for pattern analysis: which blocks are structurally misused, which loading regulations are routinely ignored, and where dynamic curbside pricing or physical redesign would yield the highest return.
The operational outcome is a continuously updated curbside utilisation layer that feeds directly into traffic-management systems, parking policy reviews, and infrastructure capital programmes. Municipal governments that control this data can set pricing, enforce regulations, negotiate with delivery platforms from a position of evidence, and protect residents' personal movement data under their own legal framework — none of which is possible when the intelligence is rented from a foreign commercial operator.
Frequently asked
What exactly does a satellite see at curbside resolution, and how is it different from traffic cameras?
A 0.5 m GSD satellite image can distinguish car-sized objects, loading bays, bike-share docking stations, and even pavement markings. Unlike fixed traffic cameras, a satellite covers an entire city grid in a single pass without requiring installation, maintenance, or power infrastructure. The trade-off is that satellites revisit a given point intermittently rather than continuously, so they complement rather than replace ground sensors for real-time enforcement.
Can a small nation afford to build this capability rather than buying imagery from Planet or Maxar?
A sovereign 4–8 microsatellite constellation optimised for 1 m optical urban imaging can be built and launched for roughly $20–40 million depending on orbit and ground segment choices — comparable to three or four years of high-volume commercial imagery licensing from a tier-1 provider. ESA's NewSpace benchmarking data and World Bank Smart Cities financing instruments both recognise this as a viable entry point for middle-income countries. Sovereignty dividends — uninterrupted access, no licence restrictions, domestic data residency — compound the economic case over a 10-year horizon.
How does curbside mapping connect to revenue generation for a city government?
Accurate kerbside inventory underpins dynamic pricing of loading zones, parking permits, and ride-hail drop-off fees. Transport for London's evidence base shows that cities recovering curbside-management costs through dynamic pricing can generate £15–30 million per year in medium-sized urban cores. Satellite-derived ground-truth also strengthens enforcement prosecutions by providing time-stamped, court-admissible imagery of illegal dwell events.
What is the minimum revisit frequency needed for curbside policy to be useful?
For strategic planning — mapping the citywide distribution of loading bays, informal stops, and kerb-use conflicts — a 24-hour revisit cycle is adequate. For operational enforcement support, 4–6 passes per day provides statistically significant occupancy sampling. True real-time enforcement still requires ground sensors or cameras; satellite imagery fills the gap for areas where sensor deployment is economically impractical.
How do AI models turn raw satellite pixels into curbside analytics?
Computer vision pipelines use convolutional neural networks trained on labelled overhead imagery to detect, classify, and count kerbside objects — parked vehicles, motorcycles, delivery cargo, construction skips — and compare them against a basemap of permitted uses per kerb segment. Change-detection algorithms then flag new conflicts between observed and permitted states. Open frameworks such as the OGC API – Features standard enable these outputs to flow directly into city GIS platforms.
What privacy safeguards are required when imaging public streets from orbit?
Most jurisdictions require that imagery used for enforcement purposes be processed to remove or blur personal identifiers — licence plates and faces — before storage or sharing; this is codified under GDPR Article 5 in the EU and analogous national laws elsewhere. A sovereign constellation operator can bake these anonymisation steps into the ground-segment processing pipeline, with audit logs that satisfy national data-protection authorities without exposing raw imagery to a foreign commercial vendor.
How does curbside mapping interact with autonomous vehicle deployment planning?
Autonomous vehicles depend on high-definition maps that include accurate, up-to-date kerbside geometry — pick-up zones, no-stopping areas, loading restrictions. A city that continuously refreshes its kerbside map from satellite imagery can publish those updates via OGC-compliant APIs, giving AV operators a live ground-truth layer without requiring individual operators to resurvey streets. This positions the sovereign city as the authoritative data custodian rather than ceding that role to a private mapping company.
What happens to historical curbside data under a sovereign model versus a commercial subscription?
Under a commercial subscription, historical archives are typically licensed per-scene and can be withdrawn, repriced, or restricted at the vendor's discretion; some contracts explicitly prohibit use of archived data after licence expiry. A sovereign constellation accumulates an unencumbered national archive, which supports longitudinal studies of urban change, legal proceedings, and infrastructure-investment evidence bases — all of which have compounding value that is impossible to replicate retrospectively if the data was never collected.