A method that produces a plausible result has not automatically demonstrated that it works for its intended users. Validation requires a defined question, suitable reference evidence, and an evaluation design that reflects how the method will be used. This perspective examines which observations should be compared, how they should be sampled, and which errors matter in practice. It also considers the difference between evidence used to develop a method and evidence reserved for evaluating it. The objective is a test that can reveal weaknesses as well as confirm useful performance.
This working guide considers viewing geometry and brdf effects within urban green-space mosaics. The immediate decision is to select green-space locations for assessment. Useful reference evidence includes canopy cover, surface context and appropriately timed measurements. The case is an illustrative planning context, with assumptions to check against the actual sensor, landscape and project requirements.
The setting: Urban green-space mosaics
Urban green spaces form mosaics of trees, grass, bare ground, paths, buildings and water. The observation needed depends on whether the project concerns vegetation cover, land-surface temperature or a particular use of the space. These are distinct variables. Visible and near-infrared imagery can describe vegetation-related patterns, while a temperature study requires suitable thermal measurements and an interpretation that respects what those measurements represent rather than treating greenness as a temperature sensor.
Define park or management boundaries and inspect how surrounding urban surfaces contribute to each observation. Shade, seasonal foliage and maintenance can affect vegetation comparisons, while acquisition timing matters strongly to a thermal interpretation. Data from different instruments may also have different spatial support. A fine-looking display should not suggest that a coarse thermal measurement resolves the same detail as an optical image or an individual tree inventory.
Field validation should record the target variable with methods and timing suited to the intended claim. Canopy observations, surface measurements and information about how people use a site provide different evidence streams. A remotely observed pattern may help choose assessment locations without establishing a complete account of human exposure or park performance. Evaluate the specific management decision, and communicate the scale, observation time and uncertainty that determine how the resulting map can be used.
The observed surface has a direction-dependent response
A surface does not necessarily reflect incoming light equally in every direction. Its measured brightness can change as the positions of the Sun, surface and sensor change, even when the surface itself is similar. The bidirectional reflectance distribution function provides a framework for describing this directional behaviour. In practical image analysis, the relevant lesson is that a difference between observations may contain a geometry effect as well as a real surface change. Canopy structure, shadows and surface roughness can all influence that relationship.
The magnitude and pattern of a geometry effect depend on the landscape and spectral measurement. A row crop, closed forest and exposed soil need not behave in the same way. Terrain introduces another layer because slope and aspect alter local illumination. A simple brightness adjustment is not automatically an adequate solution. A workflow should first identify the type of comparison it needs and whether the available product already includes a relevant normalisation. Applying an additional correction without understanding prior processing can create a new inconsistency.
Normalise geometry with a clear purpose
Geometry normalisation uses a model or adjustment to express observations under a more consistent reference condition. It is an approximation designed for a defined purpose, and its performance should be checked rather than assumed. Keep the input geometry, model assumptions and processing version where they are available. Compare representative surfaces before and after the adjustment. A useful test examines whether the correction reduces systematic variation in stable areas while preserving changes that are supported independently by management records or field observations.
For an agricultural service, the importance of this effect depends on the size of the change being interpreted. A subtle trend can be more sensitive to residual geometry variation than a major harvest event. Evaluate whether an apparent improvement in a time series survives across seasons and fields, including those with different canopy arrangements. Retain a clear explanation of what has been normalised and what remains uncertain. Viewing geometry becomes manageable when it is treated as part of the measurement context instead of being attributed silently to crop condition or land-cover change.
Deliver quality flags that users can interpret
Quality information should describe the condition of an observation in terms that remain useful through the workflow. A single valid or invalid label may be sufficient for some calculations, but operational users often need to know why an observation was qualified. Distinguish acquisition limitations, processing failures, uncertain reference matches, and conditions that make interpretation ambiguous. These categories point towards different responses and should not be merged without a clear purpose.
Define a small, documented set of flags with stable meanings. Record whether a flag applies to a complete scene, a geographic unit, an individual measurement, or a derived result. A unit containing a mixture of qualified and unqualified observations may also require a coverage measure. Include the rule used to combine flags when summarising data so that a favourable average does not conceal a substantial fraction of unusable input.
Test the flags with the people receiving the product. Ask what they would do differently for each state and whether the terminology supports that choice. Keep the original flags available even if the interface presents a simpler summary. When flag definitions change, version the rules and review their effects on historical comparisons. Useful quality communication makes the condition of the evidence visible at the moment when a user decides how much confidence to place in the result.
Design an evaluation that reflects deployment
A model evaluation should resemble the conditions in which the model will be used. Start by identifying which relationships the development data share: nearby locations, repeated observations of the same unit, common collection teams, or measurements from a single period. These relationships can make records less independent than they appear. An evaluation that ignores them may test the ability to recognise familiar conditions rather than the ability to operate in a new setting.
Choose a split strategy that matches the intended transfer. Holding out locations, seasons, operational units, or acquisition conditions can answer different questions. Document which question each split is designed to address, and keep the held-out evidence separate while selecting features and tuning the method. Repeatedly inspecting the same evaluation set to improve a model gradually makes that set part of development, even if it retains its original label.
Report results at a scale that is meaningful to the user. A strong average can hide poor performance for important subgroups or unusual conditions. Examine error patterns and retain examples of both successful and unsuccessful predictions. Where the evidence covers only a limited range of conditions, describe that range explicitly. The evaluation then becomes a practical guide to suitable use, further testing, and the collection of additional reference observations.
Bring the evidence back to the decision
The validation record should explain the population represented by the test, the reference observations used, and the limitations of the comparison. Report meaningful errors alongside useful successes, and retain examples that challenge the method. Treat wider deployment as a new opportunity to evaluate transfer across relevant conditions. A well-designed validation programme gives users a defensible basis for deciding where the method is useful, where additional evidence is needed, and where another approach may be appropriate.
Sources & further reading
Explore the bhumisat spacecraft concept and our vision towards 2030, or return to the journal.



