Airborne & Multispectral Imaging

Deck scope. Two big themes:

1. Terminology of aerial photographs — vantage point, flight height, scale, view angle, ground coverage.
2. Displaying multispectral images — panchromatic vs multispectral; true-color vs false-color composites; additive vs subtractive color.

Five properties to describe any aerial photograph. Know the definitions:

• Vantage point — camera orientation (vertical vs oblique).
• Flight height (altitude) — how high above the ground.
• Scale — ratio of photo distance to ground distance; depends on focal length and altitude.
• View angle — field of view of the camera; depends on focal length.
• Ground coverage — actual area on the ground per photo; depends on altitude and view angle.

Camera = passive optical sensor. Uses a lens (B) to focus radiation from a scene (A) onto the focal plane (C) where the image is formed.

• “Passive” — relies on natural illumination (typically sunlight).
• “Optical” — operates in the visible and near-visible parts of the spectrum.

Vantage point = camera orientation during acquisition. Two basic categories:

• Vertical — optical axis points straight down (within 3° of vertical).
• Oblique — optical axis tilted more than 3° from vertical.

Vertical aerial photography. The camera’s optical axis is within 3° of vertical.

• Less geometric distortion than oblique — nearly uniform scale across the photo.
• Uses: planimetric maps, topographic maps, orthophotomaps, and DEMs (digital elevation models) — all derived via photogrammetry (stereo overlap + parallax).

Orthophotomap example. An orthophotomap is a vertical aerial photo that has been geometrically corrected — scale is uniform everywhere, so it can be measured like a map.

• Produced from vertical photography (slide 5) plus a DEM to remove relief displacement.
• Example shown: Washington D.C. downtown.

Oblique aerial photography. The camera’s optical axis deviates more than 3° from vertical.

• More geometric distortion — scale varies from foreground (large) to background (small).
• Strengths: covers a very large area in a single image; depicts terrain relief and scale in a way vertical photos can’t.
• Used for reconnaissance, scenic/interpretive imagery, and rapid situational assessment.

Topographic map example. A topographic map represents terrain using contour lines (elevation), symbols, and color. Not itself an aerial photograph — but topographic maps are frequently compiled from vertical aerial photography using photogrammetric methods.

• Shows named features (peaks, lakes, drainages) with elevation context.
• Copyright Pearson Prentice Hall / similar reference maps.

Oblique ground-level panorama. Illustrates how an oblique view captures terrain relief and perspective — peaks, ridges, valleys — in one frame, in ways a vertical nadir photo can’t.

• Labels visible: The Saddle, Hagues Peak, Mummy Mountain, Roaring River, Mount Tileston.

Flight height (altitude above ground level, AGL). For a given focal length (f):

• Higher altitude (H₂) → larger area covered, smaller scale per photo.

• Lower altitude (H₁) → smaller area, larger scale — finer detail.

• Trade-off: more area per shot means fewer shots to cover a region, but coarser detail.

Photo scale. Scale is a ratio of distances:

S = distance on photograph / distance on the ground

For an aerial photograph, this works out to:

S = f / H

• f — camera focal length.
• H — flying height above the terrain.

Scale can be expressed three ways: - Verbal — “1 cm equals 1 km”. - Representative fraction / ratio — 1/100 000 or 1:100 000. - Graphic (bar) scale — a labeled bar on the map.

• Rule: larger RF denominator → smaller scale (a 1:100 000 photo is smaller scale than 1:10 000).

View angle = field of view (FOV) of the camera.

• Given the same flight height, view angle is determined by focal length.
• Smaller focal length → larger view angle (wide-angle lens).

• Larger focal length → smaller view angle (telephoto).

• Diagram relationships: f₁ > f₂ > f₃ ⟺ α₁ < α₂ < α₃.

Ground coverage. The actual area on the ground that one aerial photograph covers.

• Determined by altitude (H) and camera viewing angle (α).
• Larger H or larger α → larger ground coverage.
• Larger coverage per frame means fewer photos needed to cover a region — but each photo sees finer detail at the center and coarser/distorted detail at the edges.

Aerial photography on film — advantages and disadvantages.

Advantages - High ground resolution (a frame of aerial film can resolve very fine detail). - Flexibility — mission-on-demand, varied altitude/lens combinations. - High geometric reliability — a stable imaging plane + known lens geometry. - Relatively inexpensive vs. satellite tasking.

Disadvantages - Daylight only — typical acceptable window is 10:00 AM – 2:00 PM local (for good sun angles). - Poorer contrast at shorter wavelengths (blue scatters in the atmosphere). - Film is non-reusable, expensive to process, storage-heavy. - Inconvenient — must fly, land, develop, scan. - Inefficient for digital analysis — has to be digitized before any modern GIS/classification work.

Panchromatic imaging. A single wide-band channel — the sensor is sensitive to radiation across a broad wavelength range.

• When that range coincides with the visible spectrum, the image resembles a black-and-white photograph.
• Advantages: high spatial resolution (no band splits), good geometric accuracy — basis of pan-sharpening workflows.
• Examples: Landsat 8/9 Band 8 (15 m pan), SPOT HRV pan (10 m), IKONOS pan (1 m).

Multispectral imaging. Several layers per pixel — one channel for each wavelength band.

• Each channel samples a different wavelength range, achieved with different filters over the detectors.
• Stored as a stack: each pixel is a vector of values (one per band).
• Enables band math (ratios, indices like NDVI) and classification.
• Example color film split: 0.4–0.5 µm blue, 0.5–0.6 µm green, 0.6–0.7 µm red.

Displaying bands separately. Any single band of a multispectral image is a grayscale image, with each pixel holding one number (a brightness value).

• A “blue band” image is not colored blue — it’s grayscale showing the brightness of the blue wavelength at each pixel. The same is true for green and red bands.
• Color appears only when we combine three bands (next slide).

Composite images. Combine three bands into a color display by assigning each band to one of the monitor’s red, green, blue (R-G-B) channels and adding them.

• The display mixes the three channels additively to produce color.
• Which three bands you pick, and which display channel you send each to, defines whether the composite looks “true color” (natural) or “false color” (intentionally remapped).

Color theory — two mixing models.

• Additive primaries: Red, Green, Blue (RGB). Used by light-emitting displays — monitors, projectors, sensors. R + G + B = White.

• Subtractive primaries: Cyan, Magenta, Yellow (CMY). Used by inks and dyes — printing, color photography. C + M + Y = Black (in theory).

• Displaying a digital image uses additive mixing (RGB).
• Color photography (film prints) uses subtractive mixing of complementary dyes.

True-color composite. Bands are assigned to the R / G / B display guns in the same order as the real-world colors — red → red gun, green → green gun, blue → blue gun — so the image looks roughly like what your eye would see.

• Landsat TM band mapping for true color: R = Band 3 (Red), G = Band 2 (Green), B = Band 1 (Blue).
• Landsat TM band numbering cheat sheet: 1-Blue, 2-Green, 3-Red, 4-NIR, 5-SWIR, 7-MIR, 6-Thermal (note 7 lies spectrally between 5 and 6 but is numbered out of order historically).
• On Landsat 8/9 (OLI): the same true-color recipe is R = B4, G = B3, B = B2 (band numbers shifted by one because OLI added Band 1 coastal/aerosol).

False-color composite (color-infrared). Shows colors that don’t correspond to real-world colors — specifically, assigns the near-IR band to the red display gun so that healthy vegetation (high NIR) glows bright red.

• Classic CIR recipe on Landsat TM: R = Band 4 (NIR), G = Band 3 (Red), B = Band 2 (Green).
• Makes vegetation, water, urban fabric much easier to discriminate than in true color.
• Uses: crop stress, burn scars, wetland delineation, timber mapping.
• On Landsat 8/9 (OLI): the equivalent CIR is R = B5, G = B4, B = B3.

False color composite — schematic repeated. The display gun mapping again: 4 → Red gun, 3 → Green gun, 2 → Blue gun.

• Mnemonic: “NIR goes red, red goes green, green goes blue.” (Each band shifts one slot toward the longer-wavelength display primary.)
• Useful check when you open imagery: if vegetation is bright red, you’re looking at a false-color (CIR) composite.