Smart Cities / Delivered project

Urban Infrastructure Intelligence Platform

A repeatable framework for turning aerial imagery into standardized evidence about roads, crossings, public assets, and change.

Explore the page ↓

Urban infrastructure mapped through restrained computer vision overlays — Smart Cities · Data → Intelligence → ActionPhoto by Denys Nevozhai · Unsplash License ↗

01 / Problem statementDelivered project · Smart Cities

The operating problem

Municipal asset inventories age quickly, and field surveys cannot economically maintain consistent coverage across rapidly changing areas.

Conditions the system must survive03 operating constraints

Municipal imagery is captured at different dates, angles, and resolutions

Asset definitions vary across departments and districts

Change detection must distinguish real change from capture variation

02 / How we approached it

From operating uncertainty to testable evidence.

The work was decomposed into four engineering decisions. Each one produced an artifact the customer could inspect, test, and carry into deployment.

01Frame

Align imagery, boundaries, and asset definitions

Evidence produced

City asset, district, and change taxonomy

02Evaluate

Detect road and public-realm features

Evidence produced

Imagery normalization and asset benchmark across districts

03Engineer

Normalize outputs into a reviewable city index

Evidence produced

Change-detection precision study by capture condition

04Operationalize

Track changes by location and acquisition date

Evidence produced

Reviewable city index with field-verification queues

03 / Deployment record

Deployed around the workflow—not beside it.

The system boundary includes where inference runs, how evidence reaches existing tools, and how people handle uncertainty after launch.

Implementation statusDelivered project

Topology

Cloud geospatial pipeline for imagery alignment, feature extraction, comparison, and versioned map publication.

Integration

Aerial imagery, district boundaries, asset registers, work orders, and GIS layers align to shared identifiers.

Operation

Each acquisition creates a versioned index; uncertain changes route to GIS or field teams before register updates.

04 / Evaluation metrics

What must be measured before the system earns trust.

Evaluation covers model behavior, workflow burden, and production performance. The metric defines the gate; the customer baseline and acceptance threshold define the target.

01Evaluation gate

Asset detection F1

How it is measured

Precision and recall by asset class and district.

What it decides

Sets which inventory classes can be updated from imagery.

02Evaluation gate

Change precision

How it is measured

Verified real changes among surfaced candidates.

What it decides

Controls field-verification workload.

03Evaluation gate

Coverage completeness

How it is measured

District area with comparable, usable, current imagery.

What it decides

Separates capture gaps from asset gaps.

04Evaluation gate

Verification efficiency

How it is measured

Field or GIS review time per accepted update.

What it decides

Measures improvement over broad manual surveys.

05 / Customer perspective

Value has to appear in the customer’s operating day.

What matters in practice

The index should prioritize where field verification matters and show the imagery and acquisition date behind every mapped change.

01Observable value signalMore current infrastructure inventories

02Observable value signalPrioritized field verification

03Observable value signalComparable evidence across districts

04Observable value signalBetter planning inputs

06 / Technology context

Tools follow the system—not the other way around.

Final architecture depends on data quality, operating conditions, integrations, risk, and evaluation criteria established during discovery.

Geospatial AIComputer visionChange detectionGISData pipelines

Test the operating assumption

Define the evidence required to move from possibility to production.

Discuss this use case ↗

Urban Infrastructure Intelligence Platform

Municipal asset inventories age quickly, and field surveys cannot economically maintain consistent coverage across rapidly changing areas.

From operating uncertainty to testable evidence.

Align imagery, boundaries, and asset definitions

Detect road and public-realm features

Normalize outputs into a reviewable city index

Track changes by location and acquisition date

Deployed around the workflow—not beside it.

What must be measured before the system earns trust.

Asset detection F1

Change precision

Coverage completeness

Verification efficiency

Value has to appear in the customer’s operating day.

Tools follow the system—not the other way around.

Related operational systems.

Privacy-aware traffic and incident intelligence

Edge AI for predictive maintenance and leak detection

Define the evidence required to move from possibility to production.