Products · Intelligent Payloads and Accessories

Multispectral Payload Integration Assessment

Multispectral sensors are calibrated instruments. The integration question is whether that calibration survives the flight and the mounting.

4 of 10Method note

Overview

Check whether a reflectance calibration panel is part of the workflow and who carries it to the field.
Band alignment between sensors must be verified after mounting, not assumed from the datasheet.
Sun angle changes reflectance between passes; for time-series work, normalise or fly within a fixed time window.
Storage and transfer volume grows quickly with band count; plan the ground segment accordingly.

Multispectral measures reflectance in bands, not a picture

A multispectral payload captures reflected energy in a small number of discrete wavelength bands, typically including some outside the visible range. The output is not primarily a picture to look at; it is a set of reflectance values per band that can be combined into indices correlated with plant properties — vigour, chlorophyll content, water stress, and in some cases species discrimination.

That framing sets the expectation correctly. Multispectral does not produce a prettier image and it does not replace either a mapping camera or a thermal sensor. It answers questions about vegetation condition that neither of the other two can, and it is worth its mass and cost only where those questions are part of the deliverable.

The corollary is that multispectral is a measurement, and like any measurement it is only meaningful under controlled conditions. Sun angle, cloud cover, atmospheric state and the sensor's own calibration all affect reflectance values. Comparing values acquired under different conditions produces differences that are artefacts rather than findings.

Bands, indices and specifying against the question

Different vegetation questions are answered by different band combinations. Vigour and biomass indices typically use the contrast between red and near-infrared reflectance. Water stress questions use bands in the short-wave infrared. Chlorophyll and nitrogen status use specific narrow bands. Selecting a sensor without specifying which of these questions the programme has is selecting a sensor without a requirement.

So write the requirement as the index or the property you need to derive, then confirm the sensor covers the bands it depends on. A sensor with more bands is not automatically more useful; what matters is whether it has the bands your index needs, with adequate radiometric quality in those bands.

Radiometric quality is the part that is least visible in specifications and most consequential in results. Consistency between bands, linearity of response, and stability over time determine whether derived indices are reliable. Two sensors with the same nominal bands can produce indices that differ enough to matter, and that difference is radiometric rather than spectral.

Calibration and the conditions discipline

Multispectral data requires radiometric calibration to convert raw values into reflectance. The common approach uses a reference panel of known reflectance captured under the same illumination as the flight, and it requires that capture to happen properly — panel clean, level, and imaged at the right time relative to the flight.

This is a field procedure with real discipline attached, and programmes that skip it or perform it inconsistently produce data that cannot be compared between sorties. Where a programme will compare across time — season-long monitoring, treatment comparison — calibration discipline is the single biggest determinant of whether the comparison is valid.

Illumination conditions deserve an explicit rule. Flights should be scheduled near solar noon where possible, under clear or consistently overcast sky, and within a defined window. Mixed illumination within or between sorties introduces variation that dominates the signal being measured. Write the rule, and be prepared to re-fly rather than to accept data acquired outside it.

Integration and the practical checks

The four standard integration checks apply: mounting provisions, available power, control and data interface, and mass and centre-of-gravity envelope. Multispectral payloads are typically light and modest in power demand, so integration is usually straightforward where the interface is documented.

Two points deserve specific attention. The first is the illumination sensor: many multispectral systems include a downward-looking light sensor to record illumination during flight, and it must have an unobstructed view. Mounting that blocks it invalidates the calibration approach. The second is synchronisation — the imagery and the illumination record must be time-aligned for the correction to work.

As with any payload, confirm the interface rather than assuming it. Our payload interface and platform compatibility note covers the checks, and where a sensor must be combined with others on one aircraft, multi-payload coordination covers the coordination work.

Assessing for your programme

The assessment sequence is: state the property or index the programme needs; confirm the sensor has the bands with adequate radiometric quality; confirm the calibration procedure and who performs it; set the illumination rule; then confirm the four integration properties.

Where the vegetation question is not part of the deliverable, multispectral is weight and cost without return, and we will say so. Where it is, the sensor is the easy part — the calibration and conditions discipline is what determines whether the data supports a conclusion.

Typical applications include aquatic vegetation change and wetland vegetation survey, where the property being measured is vegetation condition over time. Send your crop or vegetation type, the property you need, your flight window and your platform to our engineering team, and we will assess whether multispectral is warranted and what it requires.