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 writing requirements from an earth observation need in potato growing fields. The immediate decision is to focus scouting within potato fields. Useful reference evidence includes emergence records, canopy stage and representative plot observations. The case is an illustrative planning context, with assumptions to check against the actual sensor, landscape and project requirements.
Translate a user decision into measurable behavior
A mission requirement should connect to a decision that someone needs to make with the observations. Begin by defining the target phenomenon, geographic area, relevant season and consequence of delay or error. A request to monitor crops is too broad to guide a spacecraft design. The useful question is which crop condition must be distinguished, at what field scale, how often, and with what confidence.
Translate these needs into product requirements before allocating spacecraft specifications. Consider spatial detail, spectral information, radiometric performance, location accuracy, usable revisit and delivery latency. Geometric revisit describes observation opportunities; clouds and acquisition constraints can make usable observations much less frequent. Define terms carefully so that a requirement for delivered, cloud screened coverage cannot be satisfied merely by showing an orbit access calculation.
Keep traceability through design and verification
Write individual requirements that can be checked using a stated method and acceptance criterion. Identify the conditions under which each applies, such as target brightness, viewing angle or operating temperature. Avoid combining several unrelated promises in one sentence. Retain a rationale and a link to the higher level need, allowing reviewers to understand why the requirement exists and what would change if it were relaxed.
Allocate performance across the payload, spacecraft, ground segment and processing chain. Image sharpness may depend on optics, detector sampling, motion and resampling. Delivery time may depend on acquisition planning, onboard storage, ground contacts and processing queues. Budgets make these contributions visible and expose conflicts between objectives. Preserve margin where uncertainty remains, and record assumptions rather than hiding them inside an optimistic headline specification.
Plan verification while writing the requirement. Inspection, analysis, demonstration and testing each provide different kinds of evidence. Define the test configuration, reference measurement, uncertainty and pass criterion early enough to influence the design. Verification asks whether the implemented system meets its requirements; validation asks whether that system serves the intended use. Representative users and realistic application examples can reveal an inadequate requirement even when every hardware test passes. Manage revisions through a controlled record that describes their consequences for dependent requirements, budgets and tests. This turns requirements into a working design tool rather than a document assembled only before a review.
The setting: Potato growing fields
Potato monitoring begins with the distinction between the canopy visible from above and the crop component ultimately harvested below ground. Establishment timing, variety, row structure and management can create different canopy patterns across fields. An image service should specify whether it is supporting crop-stage assessment, identifying areas for inspection or estimating a defined production variable. These objectives require different reference measurements and should not be combined into one unqualified vegetation score.
Maintain a field timeline that includes planting, emergence and the operations that can affect canopy appearance. Early observations may contain a large soil contribution, while later observations may show a more continuous canopy. Cloud and shadow contamination, field edges and neighbouring land cover can alter summaries at either stage. Compare observations using suitable quality rules and retain the acquisition dates so that a recent-looking map does not conceal an old measurement.
A field team should collect consistent crop-stage and canopy observations across both flagged and typical areas. Record management context and sampling support rather than relying on an isolated photograph. A vegetation change cannot establish the presence of a specific disease, and a canopy index is not a direct measurement of tuber yield. Evaluate any production model with appropriate harvest data and independent assessment, while keeping the simpler inspection-priority product transparent about what it actually observes.
Use observation labels with consistent meanings
A reference label should describe an observation that another person could recognise or reproduce. Broad terms such as good, damaged, or unusual can conceal substantial differences between collection teams. Begin by defining the feature or condition being recorded, the observation unit, and the evidence needed to assign a label. Where possible, distinguish a direct measurement from an interpretation made by the observer.
Provide a short collection guide with examples, boundary cases, and a way to record uncertainty. Train observers using shared cases before relying on independently collected labels. If two observers disagree, preserve the disagreement long enough to understand its cause. It may reveal an unclear definition, a genuinely ambiguous condition, a difference in observation scale, or a problem in matching the reference record to the remote measurement.
Retain the context that makes the label interpretable. Location, time, collection method, observer role, and relevant surface conditions can be essential for later review. Changes to the label scheme should be documented and related to earlier versions where a defensible correspondence exists. Consistent labels do not eliminate uncertainty, but they make it possible to investigate disagreement, evaluate analytical methods, and identify where better definitions or additional measurements would improve the evidence.
Define a representative reference sample
A reference sample should represent the situations in which a method will be used. Begin by defining the target population: locations, time periods, management conditions, surface types, or instrument states that belong within the intended application. A collection of convenient observations may be valuable for exploration, but it does not automatically support conclusions about that wider population. Make the scope of the sample visible before calculating a summary of performance.
Consider a stratified design when important conditions are unevenly represented. Relevant strata might include different landscape types, developmental stages, observation quality levels, or ranges of the variable being estimated. Choose sample locations within those groups using a documented rule, and retain the information needed to understand their selection. Where access restrictions affect the sample, record the excluded conditions so that the limits of the resulting evaluation remain clear.
Keep development and evaluation evidence separate where the purpose of the study requires an independent test. Nearby observations or repeated measurements can share information, so a random split of individual records may overstate independence. A split organised by location, period, or operational unit can provide a more demanding comparison. The appropriate design follows the intended deployment, the structure of the observations, and the errors that matter to the user.
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.



