Remote sensing

Remote sensing is, in the U.S. Geological Survey’s definition, “the process of detecting and monitoring the physical characteristics of an area by measuring its reflected and emitted radiation at a distance”. Two terms recur throughout:

1. The sensor (instrument) measures the radiation, the electromagnetic radiation (EMR). The most common sensor is a multispectral radiometer, which measures EMR at several wavelengths at the same time.
2. The platform carries the sensor over the target: a drone, an aircraft or a satellite.

These notes deal mainly with multispectral radiometers on satellites, and with how their measurements of the Earth’s surface are used to understand the physical environment. Data from them are free and plentiful: Sentinel-2 alone images every place in Denmark every few days at 10 m resolution. Handling that amount of data is part of the work. These notes use the Copernicus Browser to find an image and cut out the study area, and QGIS to analyse it.

Three kinds of sensor

• Reflected sunlight. Most sensors measure sunlight reflected by the surface, in visible and infrared wavelengths. Sentinel-2’s MultiSpectral Instrument (MSI) and Landsat 8’s Operational Land Imager (OLI) are of this kind.
• Emitted radiation. Some measure radiation emitted by the surface itself: thermal infrared, from which surface temperature can be calculated (Landsat 8’s Thermal Infrared Sensor, TIRS), or the light of cities at night (VIIRS on the Suomi NPP satellite).
• Active sensors send out their own radiation and measure what comes back: radar from satellites (Sentinel-1), lidar from aircraft and drones.

The seven elements of remote sensing

1. Energy source or illumination: the sun, or the radar or lidar on the platform.
2. The atmosphere: the radiation passes through it on its way to the target, and again on its way from the target to a satellite sensor.
3. Interaction with the target: part of the radiation is absorbed, a little is transmitted, and part is reflected, depending on both the surface and the wavelength.
4. Recording by the sensor: the sensor records the reflected or emitted radiation.
5. Transmission, reception and processing: the recording is sent to a ground station and processed into images, often together with derived products that make analysis easier.
6. Interpretation and analysis: the image is interpreted, visually or digitally, to extract information about the target.
7. Application: the information is used to understand the target better, reveal something new, or help solve a problem.

Chapters

1. Electromagnetic radiation and the atmosphere: wavelengths, the spectrum, sunlight, reflectance, scattering and absorption.
2. Earth observation platforms and sensors: satellites and their orbits, aircraft, handheld spectrometers, radar.
3. Digital images and reflectance: pixels, digital numbers, calibration, top-of-atmosphere and surface reflectance, bands.
4. Visualising multispectral images: grey-scale and colour composites, contrast stretch.
5. Spectral signatures and vegetation indices: how water, soil and vegetation reflect, and NDVI.
6. Image interpretation and classification: visual interpretation, ground truth, classification and its accuracy.

In practice: Vegetation index from satellite imagery (Task), with the Recipe Vegetation index from Sentinel-2 (Recipe - QGIS) and the worked example Vegetation greenness at Ledreborg (Example).

These notes are a revision of the compendium “Earth System Science & Remote Sensing” (Esbern Holmes, 2021), written for Google Earth Engine. The chapters on radiation, the atmosphere, orbits and images draw on the tutorial Fundamentals of Remote Sensing by the Canada Centre for Mapping and Earth Observation (Natural Resources Canada), free for non-commercial use.