Geoinformation: Remote Sensing, Photogrammetry, and Geographic Information Systems Notes

Surveying and Mapping in Transition to Geoinformation

• Mapping and surveying practices transitioned in the 1990s from discipline-oriented technologies (geodesy, photogrammetry, surveying, cartography) to methodology-oriented integrated disciplines known as geoinformatics or geomatics.

• This new concept is based on the Global Navigation Satellite System (GNSS/GPS), remote sensing, digital photography for data acquisition, and Geographic Information Systems (GIS) for data manipulation and output.

• Sustainable development requires spatial information provided efficiently through these combined methodologies.

• Geodesy is defined by F.R. Helmert (1880) as the science of measurement and mapping of the earth’s surface, requiring a reference surface for positional fixing.

• Historical progression of earth shape concepts:

  • Ancient Greece (Homer, 800 bc): A disk surrounded by oceans.

  • Pythagoras (550 bc) and Aristotle (350 bc): Postulated a sphere.

  • Eratosthenes (200 bc): Measured the distance between Syene (Tropic of Cancer) and Alexandria by cart wheel, computing a spherical radius of $5909\,km$.

  • Renaissance astronomers (Copernicus, Kepler, Galileo): Postulated an ellipsoidal figure, verified by meridional arc measurements in Lapland and Peru (1736–1737).

• Local best-fitting reference ellipsoids were established based on mass irregularities in the crust and local gravity field (verified by G. Everest in the Himalayas):

  • Clarke (1880): $a = 6378249\,m$, $b = 6356515\,m$

  • Bessel: $a = 6377879\,m$, $f = 1/298.61$

  • Hayford: $a = 6378388\,m$, $f = 1/297$

  • Krassovskij: $a = 6378295\,m$, $f = 1/298.4$

• Projections used to convert ellipsoidal coordinates to mapping coordinates:

  • Stereographic projection: For circular countries (e.g., the Netherlands, New Brunswick).

  • Gauss–Krüger ($3^{\circ}$ Transverse Mercator): Applied for meridians $3\,degrees$ apart (used in Germany, South Africa).

  • Lambert Conic Conformal: For E–W elongated countries (e.g., France).

  • Universal Transverse Mercator (UTM): A $6^{\circ}$ Transverse Mercator with a principal meridian scale factor of $0.9996$.

• Vertical reference systems are based on the geoid (equipotential surface of resting oceans), typically realized via tidal benchmarks and spirit leveling.

• The World Geodetic System 1984 (WGS 84) dimensions: $a = 6378137\,m$, $f = 1/298.257223563$.

Remote Sensing Fundamentals and Electromagnetic Radiation

• Remote sensing is the identification or survey of objects by indirect means using force fields, primarily the electromagnetic spectrum.

• Basic Laws:

  • Relation between frequency ($\nu$) and wavelength ($\lambda$): $c = \lambda \cdot \nu$, where $c$ is the speed of light (approx. $300,000\,km/s$).

  • Planck’s distribution law for spectral radiance ($L_{\lambda}(T)$): $L_{\lambda}(T) = \frac{2 h c^2}{\lambda^5} \cdot \frac{1}{e^{hc/k\lambda T} - 1}$.

  • Wien’s law for radiation maximum: $T \cdot \lambda_{max} = 0.002898\,m \cdot K$.

  • A body at $6000\,K$ (Sun) has a maximum radiance at $480\,nm$ (green light).

  • The Earth ($273\,K$ to $300\,K$) peaks in the thermal range ($8\,mm$ to $14\,mm$).

• Energy–Matter Interaction:

  • Absorption and scattering diminish energy transmission. Rayleigh scattering is proportional to $\lambda^{-4}$.

  • Atmospheric transmission is only possible through specific "atmospheric windows."

  • Relationship: $\rho_{\lambda} + \tau_{\lambda} + \alpha_{\lambda} = 1$ (Reflection + Transmission + Absorption coefficients).

  • Normal reflection coefficients for green light: Coniferous forest ($1\%$), Water ($3\%$), Meadow ($7\%$), Road ($8\%$), Deciduous forest ($18\%$), Sand ($25\%$), Limestone ($60\%$), New snow ($78\%$.

• Plant Leaf Interaction: Consists of a transparent epidermis, a palisade mesophyll (reflects green, absorbs red via chlorophyll), and a spongy mesophyll (reflects near-infrared).

Imaging Sensors and Systems

• Photographic Process:

  • Silver halide particles in gelatin emulsion are converted to metallic silver. Gradation ($\gamma$) is the slope of the D–log H curve: $\tan \alpha = \frac{\Delta D}{\Delta \log H} = \gamma$.

  • CIE Principal Colors: Blue ($435.8\,nm$), Green ($546.1\,nm$), Red ($700.0\,nm$).

• Digital Imaging:

  • Charge-coupled devices (CCD) consist of metallic electrodes on a silicon semiconductor. Resolution is limited by sensor area and sample frequency.

• Aerial Survey Cameras:

  • Standard image format: $23 \times 23\,cm$. Cassette film lengths up to $120\,m$.

  • Lenses: Wide-angle ($f = 153\,mm$) is standard; Normal-angle ($f = 305\,mm$) is used for urban/forest surveys to reduce occlusions.

  • Image motion compensation ($IMC$) improves resolution by moving film forward during exposure.

  • Examples: Leica RC 30, Zeiss RMK TOP, Z/I Imaging DMC (Digital Modular Camera).

  • DMC 250: $16768 \times 14016\,pixels$ ($15\,cm$ GSD at $h = 3000\,m$).

  • Leica ALS 70 Laser Scanner: $500\,KHz$ scan rate, up to $5000\,m$ altitude.

  • Vexcel UltraCam Eagle: $20010\,pixels$ across, exchangeable lens system ($80\,mm$ or $210\,mm$).

• Radar Imaging:

  • Active microwave bands: X band ($2.4-3.8\,cm$), C band ($3.8-7.5\,cm$), L band ($15-30\,cm$).

  • Doppler effect permits Synthetic Aperture Radar (SAR) with high azimuthal resolution.

  • Radar Interferometry: Uses two antennas to observe phase difference ($\Delta\phi$) to derive digital elevation models ($DEM$): $\Delta\phi = \frac{4 \pi}{\lambda} \Delta r$.

Remote Sensing Platforms and Repeatability

• Satellite Characteristics:

  • Earth observation orbit (approx. $1000\,km$): Period $U = 105.6\,min$, Velocity $\nu = 7.34\,km/s$.

  • Geostationary orbit ($35790\,km$): Period approx. $24\,hours$.

• Major Systems:

  • Landsat 1 (1972): First resources satellite ($80\,m$ pixels).

  • Landsat 7: $15\,m$ panchromatic, $30\,m$ multispectral ($MS$).

  • Spot (France): $2.5\,m$ to $10\,m$ pixels, sideways tiltable for rapid coverage.

  • Ikonos 2: First commercial $1\,m$ satellite (1999).

  • Quickbird: $0.6\,m$ ground pixels.

  • TerraSAR-X: Radar satellite with up to $1\,m$ GSD in Spotlight mode.

• Repeatability vs. Resolution:

  • Global (Meteorology): High repeatability ($30\,min$), low resolution ($1-5\,km$).

  • Regional (Resources): Medium repeatability ($16-26\,days$), medium resolution ($10-30\,m$).

  • Local (Cartography): Low repeatability ($years$), high resolution ($0.1-1\,m$).

Photogrammetry and Geometric Restitution

• History: Nièpce and Daguerre invention of photography (1839). Laussedat (1851) built first terrestrial survey camera. Finsterwalder (1889) surveyed glaciers.

• Stereo Restitution concepts:

  • Interior Orientation: Relating image pixel coordinates to the projection center using fiducial marks.

  • Relative Orientation: Orienting one photo to another to eliminate vertical parallaxes ($py$) at six "von Gruber points."

  • Absolute Orientation: Reorienting the stereo model to ground reference using control points ($7-parameter$ transformation).

• Analytical Photogrammetry:

  • Uses collinearity equations to model the straight line between image point, perspective center, and object point.

  • Space Resection: Determining the six orientation elements ($x_{o'}, y_{o'}, z_{o'}, \omega, \varphi, \kappa$).

  • Space Intersection: Calculating 3D coordinates from two orientations and image coordinates.

• Bundle Block Adjustment: Programs like BLUH can adjust $6000$ photos and $200,000$ points simultaneously with self-calibration (additional parameters for lens distortion).

• Digital Operations:

  • Image Pyramids: Resampled overview images (e.g., $2 \times 2$, $4 \times 4$) to speed up searching.

  • Area-based Matching: Cross-correlation coefficient ($r_{ij}$) comparison.

  • Least Squares Matching: Minimizes gray level differences via geometric transformation.

  • Semiglobal Matching (SGM): Pixelwise matching with 2D smoothness constraints to improve handling of occlusions in urban areas.

• Digital Elevation Models (DEM):

  • Interpolation via Stochastic fields, Triangulated Irregular Networks (TIN), or Finite elements.

  • Orthoimages: Generated by geometric resampling of images onto the DEM.

Geographic Information Systems (GIS)

• Definition: A system for digital spatial data management (acquisition, analysis, visualization) for planning and monitoring.

• Hardware: CPUs (speed in MIPS), Storage (Terabytes), Graphics systems for 2D/3D viewing.

• Vector System Topology Relationships:

  • Line-Area Table (determines left/right adjacency).

  • Coordinate Table (all points).

  • Line-Point Table (from/to node connectivity).

• Spatial Queries: Inclusion, adjacency, intersection, and buffer zone generation.

• Database Models:

  • Relational: Most common, using tables (tuples) and primary keys.

  • Object-Relational: Extensions such as Oracle Spatial or Esri SDE for handling complex spatial data types.

• ArcGIS Levels: ArcReader (viewing), ArcView (query), ArcEditor (editing), ArcInfo (advanced processing).

• Applications:

  • Cadastral GIS: Real estate management combining maps and land registers (deeds/titles).

  • Facility Management: Infrastructure inventory (utility networks for water, gas, electricity).

  • 3D City Models: Created from footprints and image matching for antenna placement or landscape planning.

Positioning and Cost Considerations

• GNSS (NAVSTAR-GPS):

  • Carriers: L1 ($1575.42\,MHz$) and L2 ($1227.60\,MHz$).

  • Codes: P-code (Precision, $10.23\,MHz$) and C/A code (Clear/Acquisition, $1.023\,MHz$).

  • Differential GPS ($DGPS$): Uses a master and rover station to reach $1\,cm$ accuracy by eliminating ionospheric and clock errors.

• Costs (World Market approx.):

  • Aerial Photography Mobilization: $3000$ to $5000$ USD.

  • Scanning: $15$ USD per image.

  • DEM Generation: $120$ USD per image.

  • Urban 1:1000 Line Mapping: $3568$ USD/km2.

  • Satellite Radar Interferometry DEM ($\pm 12\,m$): $2$ USD/km2.

  • Ground Parcel Surveys: $5$ to $10$ USD per parcel (e.g., Albania, Georgia).

• Kondratjev’s Economic Cycles: Postulates technological inventions trigger $50-year$ periods of rapid growth (Current: Electronics/Internet/Interdependence).