Image preprocessing
Quality checks, normalisation, band handling and preparation for downstream analysis, where supported by the source.
Satellite image analytics
Turn image data into structured information.
Use image-processing and analytical workflows to extract features, classify image areas, assess patterns, and prepare data that supports a defined business or operational question.
Example visualisationCapabilities
Analytical building blocks
Distinct from finding individual objects or comparing dates, image analytics turns whole images into measurable, structured information. Each capability is used only where the source data supports it.
Feature extraction
Derive measurable image features or visual characteristics relevant to the question being asked.
Land-cover / land-use analysis
Classify broad surface categories when the sensor, resolution and methodology are suitable.
Area & count estimation
Calculate approximate areas or counts where resolution, georeferencing and validation support meaningful results.
Multispectral analysis
Use several spectral bands when the source data includes them and the use case supports it. Not every image is multispectral.
Geospatial data preparation
Organise suitable outputs into structured tables and spatial datasets ready for your own tools.
Analytical reporting
Present findings, assumptions, quality notes and limitations in a format decision-makers can follow.
From pixels to useful results
A simple, inspectable pipeline
Each stage is documented, so you can see how a result was produced and what it depends on.
- Source imagerySensor, bands, date and resolution recorded
- PreprocessingQuality checks and preparation
- Feature analysisClassification, extraction, measurement
- Quality reviewChecked against reference information
- Structured resultTables, datasets and a report
Potential applications
Where image analytics can help
Suitability depends on the question, the sensor and its bands, whether a time series is available, the resolution, and what reference information exists to check results against.
Land-cover assessment
Estimate the share of broad surface types across a defined area.
Vegetation-related analysis
Assess vegetated versus bare surfaces, where appropriate bands are available.
Selected environmental assessment
Characterise land cover around a site or within a study area.
Site-area comparison
Compare the composition of several sites using the same method.
Structured imagery datasets
Prepare consistent, analysis-ready data for your own models or GIS work.
Example outputs
Classification, summary, export, report
Illustrative outputs for the example landscape above. Percentages are computed from the areas drawn in that example image, not from a customer project.
Example visualisationArea summary by class
Illustrative data| Class | Share | Proportion bar |
|---|---|---|
| Cropland | 48.1% | |
| Grassland | 20.9% | |
| Bare soil | 12.2% | |
| Woodland | 9.5% | |
| Built-up | 7.9% | |
| Water | 1.4% |
Structured data export
CSV excerptclass,share_of_image_pct
cropland,48.1
grassland,20.9
bare_soil,12.2
woodland,9.5
built_up,7.9
water,1.4Technical report
A concise report sets out the question, the imagery used (source, date, resolution), the method, the results, quality-review findings, and the limitations to keep in mind when using the numbers.
Real-world units when supported
With properly georeferenced imagery, areas can be reported in hectares or square metres and exported as spatial datasets, rather than as shares of an image.
FAQ
Image analytics questions
Are all satellite images multispectral?
No. Some imagery only provides visible colour (RGB) or a single panchromatic band. Multispectral methods are proposed only when the source provides the bands the analysis needs.
How accurate are area estimates?
They are approximate. Accuracy depends on resolution, how clearly classes can be separated, georeferencing quality and how the output is validated. Reports state these assumptions alongside the figures.
Can you work with our existing classification scheme?
Often, yes. We check whether each of your classes can be distinguished in the available imagery and suggest merging or adjusting classes where it cannot.
Related services
Combine with other capabilities
Analytics outputs feed naturally into change comparisons or custom processing pipelines.
Discuss an image-analysis requirement
Share the question you need answered, the area of interest, and any imagery or reference data you already have.