Lecture 1: Raster and geospatial images Flashcards

Question

Image histogram

Answer 1

An image histogram shows the distribution of pixel values. Pixel values are function of their encoding. Their distribution can be expressed as a count or as frequency

Answer 2

Image acquisition = PLATFORM + INSTRUMENT(S) [≃ sensor(s)] Platform = instrument carrier. It can be spaceborne, airborne or ground-based. Sensor = (emission and) measurement of electromagnetic radiation (EMR) Instrument = sensor(s). (Instrument and sensor are often used as synonyms)

Answer 3

Measure EMR from an external source (usually, sun illumination and Earth’s thermal radiation) What do we measure? During daytime: → Sunlight reflected by the Earth’s surface → Sunlight travels through the atmosphere → some is absorbed, some scattered → The rest reaches the surface, where it is partly reflected back and partly absorbed → Reflected part travels back through the atmosphere to the satellite sensor → The atmosphere again absorbs/scatters part of it, only a fraction reaches the sensor → This reflected energy is mostly in the visible and near-infrared range At night: → We measure thermal infrared energy emitted by the Earth’s surface (heat) → This energy also passes through the atmosphere → The atmosphere emits infrared radiation too → Only some of the surface-emitted energy reaches the sensor → Requires a sensor capable of detecting thermal spectral bands Common wavelengths ▪ Ultraviolet (UV): 0.01-0.38 µm ▪ Visible (VIS): 0.38-0.75 µm ▪ Near-infrared (NIR): 0.7- 1.3 µm ▪ Shortwave infrared (SWIR): 1.3-3 µm (= “Mid-infrared”) ▪ Thermal infrared (TIR): 3-15 µm ▪ (Far infrared: 3-100 µm) Main imaging acquisition techniques: Pinhole sensor and Pushbroom sensor

Answer 4

Send its own EMR + measure the reflected signal (e.g., most satellite-based radar sensors)

Answer 5

Energy transmitted at the speed of light through oscillating electric and magnetic fields. Characterized by a wavelength (𝜆, in m) or a frequency (𝜈, in Hz) → Relationship: 𝜆 𝜈 = c(𝜈) (c (𝜈) being the speed of light for a given 𝜈)

Answer 6

Usual trade-off ▪ High spatial / Low temporal ▪ Low spatial / High temporal Low spatial / High temporal → For rapidly evolving events affecting large areas * Forest fires * Cyclones and hurricanes * Volcanic gas and ash emissions * ... Sub-daily acquisition → Real-Time Monitoring Daily acquisition → Near Real-Time (NRT) Monitoring High spatial / Low temporal → For mapping and smaller scale detection of events * Floods * Volcanic eruptions * Landslides * Earthquakes * Local climatic events * …

Answer 7

* 1 satellite * Revisit time = 16 days * Spatial resolutions: 15, 30, 100 m vs. * 2 satellites * Revisit time = 5 days * Spatial resolutions: 10, 20, 60 m

Answer 8

Planetscope ➢ Constellation of >130 cubsats (Dove) ➢ 3 or 4-band multispectral (B, G, R, NIR) ➢ Spatial resolution: 3.7 m ➢ Temporal resolution: 1-3 days ➢ Small tiles of 24.6 km x 6.4 km → Better chances to have cloud-free images

Answer 9

TROPOMI combines high spatial resolution and low detection limits ➢ ideal for SO₂ monitoring. Compared to OMI, TROPOMI clearly locates the SO₂ plume source

Answer 10

Geostationary = follow a “geosynchronous equatorial orbit” ▪ Orbital period = Earth’s rotational period → Always look at the same part of the Earth ▪ One image every 10-15 minutes ▪ High temp. / low spatial resolution ▪ Mostly weather and navigation satellites

Answer 11

▪ Daily temporal resolution ▪ Large footprint ▪ 0.5 → 2 km spatial resolution ▪ For multi-day events evolving over a large area ▪ e.g., vegetation monitoring, wildfires, cyclones, etc.

Answer 12

▪ 10-30 m spatial resolution (60-100 for TIR) ▪ 16 → 5 days temporal resolution ▪ Bands mostly in VIS, NIR and SWIR ▪ For detailed mapping and detection of small events ▪ e.g., floods, landslides, earthquake impacts, volcanic eruptions, etc.

Answer 13

▪ Spatial resolution: ➢ < 1 m in panchromatic ➢ 1-4 m in multispectral (VIS + NIR) ▪ No systematic acquisition → programmed tasks + agile sensor ▪ Rarely free (commercial satellites) ▪ (tri-)stereo capabilities ▪ For detailed and emergency mapping

Answer 14

Pinhole sensor (frame camera) → Works like a traditional camera → Captures a full image at once through a lens → Projects the scene onto a 2D sensor → Image has a fixed size but suffers from geometric distortion, especially near the edges → Distortion depends on lens properties and viewing angle

Answer 15

Pushbroom sensor (line scanner) → Works like a scanner: captures one line of the image at a time → The sensor moves forward (e.g. on a satellite) while recording successive lines → Spatial resolution across the swath depends on sensor design → Spatial resolution along the flight direction depends on satellite speed and timing → Can have distortion in one direction due to timing errors or motion

Answer 16

Sentinel-2 carries the MSI instrument, which measures reflected light in multiple spectral bands. Each band covers a specific range of wavelengths = spectral resolution. Narrow bands (small boxes) → good for detecting specific features (e.g. gases). Wide bands → cover broader ranges but with less detail. Bands are placed in regions where the atmosphere allows transmission. At wavelengths where transmission = 0, all energy is absorbed or scattered → no signal. If signal is present where it shouldn’t be (low transmission), it’s likely from clouds or noise → use cloud masking.

Lecture 1: Raster and geospatial images Flashcards

(40 cards)