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 connecting drone surveys with satellite observations around aquaculture pond clusters. The immediate decision is to prioritise pond-cluster monitoring visits. Useful reference evidence includes pond boundaries, operation status and timed water measurements. The case is an illustrative planning context, with assumptions to check against the actual sensor, landscape and project requirements.
The setting: Aquaculture pond clusters
An aquaculture landscape contains ponds at different operating stages together with embankments, channels, roads and service areas. A satellite observation may support a pond inventory or a broader review of visible water patterns, but it should not be assumed to measure conditions within every individual production unit. Begin with boundaries and a record of pond status. Empty, recently filled and active ponds can have different appearances for ordinary operational reasons.
The size of a pond relative to the observation footprint determines how much embankment or neighbouring water contributes to a measurement. Review alignment and exclude unsuitable mixed areas where necessary. Surface reflections, cloud and atmospheric conditions can complicate optical comparison, while short-term management changes may occur between usable acquisitions. Retain observation dates and quality masks so that the image record does not imply continuous knowledge of rapidly changing pond conditions.
Validation should record the unit, collection time, operational context and the measurement method for any water variable being studied. A sample from one pond cannot automatically validate an adjacent pond or an entire cluster. A colour anomaly may help select a monitoring visit without establishing a particular biological condition or production outcome. More specific estimates require a demonstrated relationship with representative measurements, while inventory and inspection-priority products can be evaluated against the practical decisions they support.
Give each platform a clear role
Drone surveys and satellites observe the same landscape at different scales and with different constraints. A drone can collect detailed local observations at a selected time, while satellites provide repeated coverage over larger regions. Combining them is useful when the local survey answers a specific question: checking crop rows, describing canopy gaps, measuring a surface, or supporting reference interpretation. Define that role before planning the flight. A detailed orthomosaic does not automatically become accurate reference data merely because its pixels are small.
Radiometric consistency requires attention to exposure settings, sensor response, illumination, and calibration targets where appropriate. Geometric quality depends on camera calibration, flight geometry, image overlap, tie points, and the accuracy of positioning and ground control. Surface reconstruction can struggle over water, moving vegetation, uniform textures, or repeated patterns. Keep independent checkpoints separate from control points used to fit the reconstruction, and evaluate horizontal and vertical accuracy against the requirements of the intended measurement.
Compare like with like
Align coordinate systems, vertical datums where relevant, and acquisition times before comparing products. A field can change after irrigation, rainfall, harvest, or a different sun angle, so a convenient date may not represent the satellite overpass. Satellite bands also have different spectral response functions from a drone camera. Matching band names does not guarantee that their reflectance values are equivalent. Describe these differences and select a comparison that the measurements can support.
Aggregate drone observations to the satellite pixel footprint or another justified analysis unit rather than simply sharpening the satellite grid. Include edge effects, positional uncertainty, and spatial heterogeneity in the comparison. A satellite pixel may mix trees, soil, shade, and roads that the drone resolves individually. Sample drone sites across the conditions present in the larger region; surveying only easy access sites can bias a calibration model. Keep training and evaluation locations independent when drone products support machine learning. Deliver the local survey with flight dates, processing settings, calibration records, accuracy checks, and known gaps. The combined workflow should make the relationship between local detail and regional inference explicit, including where the drone sample does not justify extrapolation.
Check spatial alignment before comparison
A comparison between images, channels, maps, or field measurements assumes that the observations refer to the same locations at a suitable level of precision. Small alignment errors can become especially important near boundaries or narrow features. Before interpreting differences, inspect whether identifiable, stable features agree between the sources. A common coordinate reference is necessary for many workflows, but it does not by itself demonstrate accurate alignment.
Use control features that are appropriate for the observation scale and remain stable across the comparison period. Evaluate alignment in more than one part of the area, because a single location cannot reveal spatially varying distortion. Record the method used to estimate any correction and retain independent features for checking the corrected result. Where surfaces have genuinely changed, avoid using those changes as evidence of an alignment error.
Propagate the implications of remaining misalignment into the analysis. A workflow may exclude boundary pixels, summarise at a broader scale, or qualify interpretations for small features. The choice should follow the application rather than a preference for visually smooth output. Preserve both the original and corrected geographic references, and document resampling decisions so that future users can understand how measurements were moved onto a common spatial framework.
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.



