Multispectral Imagery and Band Interpretation

Reading broad-band satellite images before jumping to indices

Published

April 4, 2026

Before You Start

You should know
That different surfaces reflect differently at different wavelengths.

You will learn
How to read common multispectral bands, why false-colour images are useful, and what broad-band patterns often signal vegetation, soil, water, and built surfaces.

Why this matters
Many higher-level products begin with a simple multispectral image. If you cannot interpret the bands, the index or classification built on top of them stays opaque.

If this gets hard, focus on…
The key habit is to ask what each band emphasizes and what contrast it creates relative to the others.

A multispectral image is not just a picture with more colours. It is a small set of carefully chosen wavelength windows, each selected because it highlights a different property of the surface. One band emphasizes chlorophyll absorption. Another emphasizes leaf structure. Another responds strongly to water content. Another separates warm roofs and asphalt from cooler vegetation. Reading a multispectral image means learning to ask what each band is sensitive to.

This chapter is the bridge between basic reflectance ideas and vegetation indices. Before computing NDVI or EVI, we should be able to look at the bands themselves and understand why those indices work.

1. The Question

How do we interpret a multispectral image before turning it into an index, classification, or map product?

Usually we start by asking:

  1. Which wavelengths are being sampled?
  2. What kinds of surfaces reflect strongly or weakly in each band?
  3. What contrast will appear if we combine those bands into an image or a ratio?

2. Common Broad-Band Regions

Multispectral optical sensors often include some version of these bands:

  • blue: sensitive to atmospheric scattering, shallow water, and some sediment or haze effects
  • green: useful for visible brightness and vegetation appearance
  • red: strongly affected by chlorophyll absorption
  • near-infrared (NIR): strongly reflected by healthy vegetation structure
  • shortwave infrared (SWIR): sensitive to water content, dryness, and some mineral differences

A Simple Surface Pattern

Three broad surface types behave very differently:

  • healthy vegetation: low red, high NIR, lower SWIR when moist
  • bare soil: moderate red, moderate NIR, often rising toward SWIR
  • water: low reflectance in NIR and SWIR

Those broad contrasts are enough to explain why so many practical remote-sensing tools work.

3. False-Colour Images

A false-colour composite assigns non-visible bands to visible display channels.

A famous example is:

  • red display channel = NIR
  • green display channel = red
  • blue display channel = green

In that display:

  • healthy vegetation often appears bright red
  • water often appears dark
  • urban surfaces often appear cyan, grey, or pale tones

The image is called “false colour” not because it is fake, but because the displayed colours are chosen to highlight useful contrasts rather than to mimic human vision.

False-Colour Logic

Band Composites Work By Reassigning What Each Display Channel Shows

A false-colour image becomes easier to read once you track which measured band is feeding the red, green, and blue channels on the screen.

Sensor bands

red band: strong chlorophyll absorption

NIR band: strong vegetation reflectance

green band: visible brightness and colour context

Display channels

screen red ← NIR

screen green ← red band

screen blue ← green band

measured bands red green NIR display RGB red green blue healthy vegetation looks red because NIR is sent to the red screen channel
False colour is a remapping of measured bands into display channels so that physically meaningful contrast becomes visually obvious.

4. A Practical Interpretation Chart

This chart is intentionally broad rather than sensor-specific. The value is in the pattern:

  • vegetation jumps high in NIR
  • water collapses in NIR and SWIR
  • soil tends to rise more steadily

5. Worked Example By Hand

Suppose one pixel has these broad-band reflectances:

  • red = 0.07
  • NIR = 0.42
  • SWIR = 0.16

That pattern suggests:

  • strong chlorophyll absorption in red
  • strong internal leaf scattering in NIR
  • moderate water-sensitive SWIR response

So the most likely interpretation is healthy vegetation with non-trivial moisture content.

Now compare another pixel:

  • red = 0.18
  • NIR = 0.20
  • SWIR = 0.28

This second pattern shows:

  • little red absorption
  • no strong NIR jump
  • stronger SWIR response

That pattern is more consistent with bare soil or dry sparse cover than with a dense healthy canopy.

6. Why This Comes Before Vegetation Indices

NDVI works because:

\text{NDVI} = \frac{\rho_{NIR} - \rho_{red}}{\rho_{NIR} + \rho_{red}}

But that ratio only makes sense if you already understand why vegetation separates so strongly between those two bands.

So a good reading order is:

  1. understand the bands
  2. understand the broad spectral contrast
  3. then compute the ratio

That way the index feels like a compressed summary of a physical pattern, not like a magic formula.

7. If This Gets Hard, Focus On

  • multispectral means a few carefully chosen broad bands
  • each band emphasizes different surface behaviour
  • false colour is a way to make useful contrast visible
  • vegetation, soil, and water differ most clearly in red, NIR, and SWIR

That foundation makes the vegetation-index chapter much easier to read and trust.