Remote sensing is the acquisition and interpretation of information about an object, area, or phenomenon without direct physical contact. In environmental science, it usually involves instruments that measure reflected or emitted electromagnetic radiation from Earth’s surface and surroundings. Sensors may operate from satellites, aircraft, uncrewed aerial vehicles, or ground-based platforms. Their observations are processed into images, maps, and estimates of physical properties. Remote sensing includes both imaging and non-imaging measurements; broader definitions also encompass acoustic techniques, such as sonar. (natural-resources.canada.ca)
Physical principles
Remote sensing depends on how signals interact with matter. Incoming radiation can be reflected, absorbed, or transmitted, while objects also emit radiation. These interactions vary with wavelength and with properties such as composition, moisture, surface roughness, and temperature. Measurements across the electromagnetic spectrum therefore reveal information that ordinary visible-light images cannot provide. Different materials can sometimes be distinguished through their characteristic spectral responses, although those responses also depend on observation conditions. (natural-resources.canada.ca)
The measured signal is not necessarily a direct reading of the target alone. Radiation passing through the atmosphere undergoes absorption and scattering, and a sensor may record contributions from both the surface and the atmosphere. Interpreting observations consequently requires knowledge of illumination, viewing geometry, atmospheric conditions, and instrument response. A bright image pixel, for example, cannot be interpreted independently of the wavelength band and acquisition conditions. (natural-resources.canada.ca)
Passive and active sensors
Passive sensors detect naturally available radiation. Visible and near-infrared instruments commonly measure sunlight reflected by the surface; thermal instruments measure emitted radiation. Reflected-sunlight observations require illumination, whereas thermal observations can be acquired at night. Clouds can obstruct surface observations in visible and infrared wavelengths. Passive microwave instruments measure naturally emitted microwave radiation and provide a different set of environmental measurements. (natural-resources.canada.ca)
Active sensors transmit a signal and measure its return. Radar uses radio waves, while lidar typically uses pulses from a laser. Lidar return times determine distances, allowing measurements to be assembled into three-dimensional point clouds representing terrain, buildings, and vegetation. Ground returns can be separated from other returns to produce elevation models. (usgs.gov)
Synthetic aperture radar combines radar observations acquired along a moving platform’s trajectory to obtain finer spatial detail than a comparable real-aperture system. Microwave radar can observe independently of sunlight and generally through cloud cover. Its response depends on wavelength, polarization, moisture, structure, and surface geometry. Radar interferometry uses phase differences between observations to investigate elevation or surface displacement. (earthdata.nasa.gov)
Platforms and historical development
The platform determines coverage, viewing geometry, and opportunities for repeated observation. Aircraft and uncrewed aerial vehicles allow flexible local surveys, while satellites provide repeated observations over extensive regions. Ground-based sensors can measure individual sites in detail and complement airborne or satellite measurements. The same platform may carry several instruments with different measurement capabilities. (usgs.gov)
The Landsat program established a long-running record of satellite land observations. Its first spacecraft, initially called the Earth Resources Technology Satellite, launched on July 23, 1972, and was later renamed Landsat 1. It carried a Return Beam Vidicon system and a Multispectral Scanner. Successive Landsat missions extended the archive, enabling comparisons of land conditions over decades, including changes associated with urbanization, wildfire, drought, and climate change. (usgs.gov)
Resolution and sampling
Remote-sensing datasets are commonly described using four dimensions of resolution:
- Spatial resolution: the size or spacing of the ground area represented by a measurement, often expressed as pixel dimensions.
- Spectral resolution: the ability to distinguish narrow wavelength intervals. Multispectral sensors record several bands; hyperspectral imaging records many narrow bands.
- Temporal resolution: how frequently observations of an area are available.
- Radiometric resolution: the ability to distinguish differences in measured signal intensity, commonly associated with digital bit depth. An 8-bit measurement has 256 possible digital values. (hideme.live)
These characteristics involve design trade-offs. Finer spatial sampling does not automatically provide finer spectral discrimination or more frequent coverage. Narrow spectral bands can distinguish subtle material differences but increase data volume. Resolution must therefore be evaluated against the phenomenon being measured rather than treated as a single ranking of image quality. (natural-resources.canada.ca)
Processing and interpretation
Raw measurements require calibration, geographic positioning, and other corrections before many scientific uses. Depending on the application, processing may address atmospheric effects, geometric distortion, or differences between acquisition dates. Image processing supports enhancement, band combinations, and extraction of useful features. Interpretation may be visual or computational, using classification and quantitative retrieval methods to translate measurements into environmental information. (natural-resources.canada.ca)
Image classification groups observations into categories such as water, vegetation, and built surfaces. Supervised learning uses labeled examples, whereas unsupervised learning identifies groupings without predefined labels. Repeated observations form time series for examining change. Remote-sensing products can also be combined with other spatial datasets in a geographic information system. (hideme.live)
Applications and limitations
Applications include tracking forest change, mapping farmland, monitoring fires and floods, observing weather, and measuring ocean temperature patterns. Satellite observations offer broad geographic coverage, while repeated measurements reveal developments that isolated surveys may miss. (usgs.gov)
Interpretation remains constrained by cloud cover, incomplete sampling, sensor characteristics, and ambiguity in the measured signal. Similar surfaces may have similar spectral responses, and one pixel may contain several materials. Field observations and independent reference data are therefore important for evaluating classifications and derived measurements. Detecting a change in imagery does not, by itself, establish its cause. (natural-resources.canada.ca)