Skip to main content

How to detect & report on asphalt patches with Birdi’s AI Model

Written by Kayley Greenland
Updated this week

Birdi’s Asphalt Patch AI model is designed to help teams quickly detect, measure, and report on asphalt patching across roadways.

This model is available on Growth and Ultimate plans.

You can run it across:

  • AI Tag – Ortho Mosaic (recommended starting point)

  • AI Tag – Map Canvas

  • AI Tag – Media (raw imagery)

Depending on your workflow, you can analyse full sites, specific sections, or individual images.

Step 1: Run Detection (Recommended: Ortho Mosaic)

For full-site analysis, start with your orthomosaic.

How to Run Detection

  1. Open your orthophoto.

  2. Select AI Tag.

  3. Choose the model:
    → Birdi Asphalt Patch Model

  4. Select your measurement type:

    • Polygon (recommended) – captures patch area

    • Marker – location only

  5. Click Detect Objects.

Why Polygon?

Polygons allow you to calculate area measurements, which are critical for:

  • Maintenance tracking

  • Cost estimation

  • Contractor verification

  • Compliance reporting

Processing time depends on the size of your dataset. Larger orthos will take longer to process.

When complete, you’ll see a new folder:

  • AI Detect - Asphalt Patch

  • Inside this folder are all detected annotations, labelled numerically.

Step 2: Review & Validate Detections

Before reporting, it’s good practice to review the outputs.

You can:

  • Select an annotation

  • Press Enter

  • Edit the polygon manually

This is useful if:

  • A patch edge needs refining

  • A large patch was partially clipped

  • Minor corrections are required

AI accelerates detection — your validation ensures accuracy.

Step 3: View All Measurements in Table Format

Once detections are complete, you can move into measurement and reporting.

Open Table View

  1. Click the AI Detect Asphalt Patch folder in the left-hand panel.

  2. On the right-hand Layers drop toolbox, select Table View.

  3. This opens a full list of:

    • All annotations

    • All measurements

    • Area values (metric or imperial depending on your workspace settings)

This gives you a complete measurable breakdown of all detected asphalt patches.

Step 4: Adjust Units (Metric / Imperial)

If you need to change your reporting units:

  1. Click the Settings icon (top right).

  2. Adjust your base units:

    • Metric

    • Imperial

The table will update automatically to reflect your selected unit system.

Step 5: Export as CSV (Finance & Quantification Workflow)

If you need a measurable output for finance, contractors, or internal reporting:

  1. In Table View, click Export

  2. Select Export as CSV

This gives you a structured file where you can:

  • Total area columns

  • Calculate cost per square metre/foot

  • Track patch quantities

  • Share with finance or asset teams

This is ideal for completing cost validation or invoicing workflows.

Alternative Detection Methods

AI Tag – Map Canvas

For smaller sections:

  • Open AI Tag on Map Canvas

  • Position the detection frame

  • Run detection on that tile only

This is useful for:

  • Quick validation

  • Partial roadway analysis

  • Faster processing

AI Tag – Media

Run the model directly on raw imagery when:

  • You want image-level validation

  • You’re reviewing edge cases

  • You need defect-specific reports

Step 6: PDF Reporting Options

In addition to CSV exports, Birdi provides two types of PDF reporting.

Option 1: Map Canvas PDF Export (One-Page Visual Report)

If you want a clean visual overview:

  1. Go to the top right toolbar.

  2. Click Export.

  3. A frame will appear on your map.

  4. Position it over the section of roadway you want to report on.

  5. Export as a one-page PDF.

This acts like a structured “print view” of your map canvas, including:

  • Orthomosaic

  • Detected polygons

  • Visible annotations

This is ideal for:

  • Council reports

  • Contractor summaries

  • Visual documentation

  • Client communication

Option 2: Media-Based PDF Report (Growth & Ultimate Only)

If you’re using AI Tag on Media to detect defects on raw imagery:

  1. Select an individual media item or the full folder.

  2. Open Table View.

  3. Click Export.

  4. Choose PDF export.

This option is available on Growth and Ultimate plans.

This report is useful when:

  • Reviewing individual defect images

  • Creating inspection documentation

  • Providing defect-by-defect visual evidence

Important Considerations

Shadows and Objects Impact Accuracy

Heavy shadows and objects can reduce detection performance.

For drone flights:

  • Fly closer to midday

  • Avoid strong directional shadowing

  • Ensure consistent lighting conditions

Cleaner imagery improves detection results.

Large Patches Being Clipped or Squared

If large patches appear partially detected:

Adjust:

  • Tile size

  • Overlap settings

Found in Advanced Settings before running detection.

Increasing overlap often improves results for larger repair areas.

Complete Asphalt Patching Workflow in Birdi

  1. Capture imagery

  2. Process orthomosaic

  3. Run Asphalt Patch AI detection

  4. Validate & refine annotations

  5. Open Table View

  6. Adjust measurement units

  7. Export CSV for quantification

  8. Export PDF for visual reporting

  9. Share outputs with stakeholders

This turns detection into measurable, shareable operational insight.

Did this answer your question?