Infrastructure & industrial sites
Visible changes to structures, storage areas or surfaces around a defined site.
Image change detection & monitoring
Understand what has changed over time.
Compare suitable images from different dates to highlight visible changes, support site reviews, and build a more consistent monitoring workflow.
Date ADate B
Example visualisation
Drag the slider, or use the arrow keys, to compare the two dates.
What can be assessed
Visible change, made easier to find and describe
Potential examples, always subject to the source imagery and the agreed analysis scope.
Built-up areas
Expansion or modification of built-up areas between suitable image dates.
Land cover
Differences in vegetation, bare ground and other surface types between dates.
Agricultural & environmental features
Changes in selected field or environmental features where the imagery supports it.
Ports & facilities
Visible changes around ports or other facilities when resolution, date and viewing geometry are suitable.
Progress reviews
Change review for a defined site, such as construction progress, where repeat imagery exists.
The monitoring workflow
How a change analysis is built
Careful preparation is what separates a real change from a difference in lighting or alignment.
-
Step 1: Define area and change
Agree the area or asset and the type of change that matters to you.
-
Step 2: Find suitable dates
Identify image dates and sources that can answer the question.
-
Step 3: Align and prepare
Co-register and preprocess imagery so the comparison is meaningful.
-
Step 4: Analyse and review
Apply change analysis and review potential artefacts and false changes.
-
Step 5: Deliver
A change layer or annotated image, a summary, or a repeat-analysis output as agreed.
Engagement models
One-off comparison or recurring monitoring
One-time comparison
Analyse two or more relevant image dates to answer a specific question — for example, what changed on a site between two inspections.
- Defined dates and area
- Change regions with a written summary
- Notes on data quality and limitations
Recurring monitoring
Repeat the analysis at an agreed cadence when new imagery is available and licensed for the intended use, using a consistent method each time.
- Consistent definitions across each update
- Comparable outputs over time
- Cadence agreed against realistic data availability
How often can imagery be updated? That depends on when imagery is acquired, the sensor's revisit, provider availability, weather and cloud for optical data, and licensing. We do not offer real-time or guaranteed daily monitoring; the cadence is agreed once data availability is confirmed.
What can cause a false change?
Not every difference is a real change
Two images of the same place almost always differ. Preprocessing and validation separate meaningful change from these effects.
Seasonal differences
Vegetation, crops and water levels vary naturally through the year.
Shadows & lighting
Different sun angles move and lengthen shadows between dates.
Cloud & haze
Optical imagery can be partly obscured or reduced in contrast.
Different sensors
Colour, band and processing differences between sources.
Different resolution
A sharper image reveals detail that was not visible before.
Misalignment
Small registration errors make edges look like new change.
Viewing angle
Off-nadir views shift tall objects and change their footprint.
Normal variation
Moving vehicles, temporary storage and everyday activity.
In the example above, Date B has a warmer ground tone and longer shadows than Date A. Those are lighting differences, so they are not marked as change.
Output examples
A change summary you can act on
An illustrative change summary for the example scene. Region IDs match the highlighted areas in the comparison above.
Change summary — Date A → Date B
Illustrative data| Region | Change type | Share of image | Review |
|---|---|---|---|
| C1 | New structure | 4.6% | Confirmed |
| C2 | New paved surface | 4.9% | Confirmed |
| C3 | Vegetation cleared | 3.1% | Confirmed |
| C4 | New access road | 2.0% | Confirmed |
| C5 | Earthworks / stockpile | 2.0% | Confirmed |
Shares are region bounding-box areas as a percentage of the example image. With georeferenced imagery, areas can be reported in square metres or hectares.
Report excerpt
Illustrative formatSummary. 5 change regions were identified between Date A and Date B. The main change is new development in the eastern plot: a new structure (C1) with an adjoining paved area (C2), connected to the road by a new access route (C4).
Notes. Vegetation was removed in the north-east (C3) and an earthworks area appeared to the south (C5). Differences in ground tone and shadow length were attributed to lighting and not reported as change.
Typical deliverables
- Before/after comparison images
- Highlighted change layer or annotated image
- Change summary table
- Short report with assumptions and limitations
FAQ
Change detection questions
Can you show changes from any date in the past?
Only if suitable archive imagery exists for that location and date, at a resolution that shows the change you care about, and under a licence that allows the use. We check availability before scoping.
Is the imagery live?
No. The analysis uses images captured on specific dates. Newly acquired imagery is used when it becomes available; we do not offer live or real-time views.
How small a change can be detected?
It depends on the image resolution and the contrast between the change and its surroundings. A change must cover enough pixels on both dates to be distinguished reliably from noise and misalignment.
Related services
Combine with other capabilities
Change analysis often builds on object detection or land-cover classification performed on each image date.
Plan a change-detection workflow
Tell us about the area or asset, the type of change that matters, and how often you would like to review it. We will check what imagery is realistically available.