OE-6.2 Digital Imaging & Sensor Resolutions

What this is and why it exists

Four resolutions describe a sensor, and the whole content of this unit is that they trade against one another.

Spatial, spectral, radiometric, temporal. No satellite maximises all four. Every mission is a chosen compromise, and knowing which compromise was chosen tells you what the data is good for.

The scanning distinction that follows is along-track against across-track. It is a design choice with real consequences for how long the sensor looks at each spot.

The vocabulary

• Spatial resolution — how small a thing on the ground one pixel covers.
• Spectral resolution — how many bands, and how narrow.
• Radiometric resolution — how finely brightness is recorded.
• Temporal resolution — how often the same place is revisited.
• Along-track scanning — a line of detectors sweeping forward with the platform.
• Across-track scanning — a mirror sweeping side to side.
• Dwell time — how long a detector looks at one spot.
• Hyperspectral — very many, very narrow bands.
• Thermal sensing — measuring emitted rather than reflected energy.

The mental model

Start with the trade, because it explains every specification you will read. Finer spatial resolution means each pixel collects energy from a smaller area, so there is less energy to work with. You can recover that by widening the bands — losing spectral resolution — or by looking longer, which costs temporal coverage. The energy arriving is finite, and the four resolutions divide it between them.

Once you hold that, a satellite specification stops being a list and becomes a statement about what the mission was for. A high-spatial, few-band sensor is built for mapping objects. A hyperspectral sensor with coarse pixels is built for identifying materials.

The two scanning designs differ in dwell time. Across-track scanning sweeps a mirror side to side, so each detector visits every position in the line briefly. Along-track scanning uses a whole line of detectors that moves forward with the platform, so each detector stares at its own column continuously. The second gets far longer per spot, which means more energy and less noise. The cost is many detectors that must all be calibrated to agree with one another.

Hyperspectral sensing takes spectral resolution to its limit: many narrow bands, giving each pixel something close to a full spectrum. That is enough to identify materials rather than merely classify surfaces, and it produces very large volumes of data.

Thermal sensing is different in kind, and this is worth being explicit about. It measures energy the surface emits rather than energy it reflects. So it works at night, and what it reports is related to temperature rather than to colour.

What you should now be able to explain or do

Name the four resolutions and explain why improving one costs another. Read a satellite specification as a statement about mission purpose. Describe along-track and across-track scanning and say which gets more dwell time. Say what along-track scanning costs in calibration. Explain what hyperspectral sensing buys and what it costs. Say why thermal sensing works at night.

Check yourself