GELE Photogrammetry & Cartography — Remote Sensing and GISSummary

Remote-sensing sensors detect reflected and emitted electromagnetic energy across specific wavelength ranges. The spectrum includes visible light (0.4–0.7 μm), near-infrared (0.7–1.3 μm), short-wave infrared (1.3–3 μm), thermal infrared (3–14 μm), and microwave (1 mm–1 m). Each material on Earth—vegetation, water, concrete, soil—has a unique spectral signature (reflectance pattern) across these bands. For example, healthy vegetation reflects strongly in the near-infrared but absorbs visible red light (used to compute vegetation indices like NDVI: Normalized Difference Vegetation Index). Understanding spectral signatures allows engineers to classify land cover, detect changes, and monitor environmental conditions across the Philippines, from Luzon agricultural lands to Mindanao forest regions.

Spatial resolution refers to the smallest ground distance that a sensor pixel can resolve—typically expressed in meters. It is determined by the sensor's instantaneous field of view (IFOV) and the altitude above ground. A 30 m pixel means each image grid cell represents a 30 m × 30 m ground area. High-resolution sensors (< 5 m, e.g., QuickBird, Pleiades) capture detailed building and parcel boundaries suitable for cadastral surveys; medium-resolution sensors (15–30 m, e.g., Landsat, Sentinel-2) suit land-cover mapping and regional planning. Low-resolution sensors (> 100 m) are used for global-scale environmental monitoring. The number of pixels across a sensor swath is calculated as: pixels = swath width ÷ spatial resolution. This parameter directly affects both data volume and mapping accuracy.

Spectral resolution is the number of discrete spectral bands a sensor records and the width (wavelength range) of each band. Panchromatic sensors record a single broad band across the visible spectrum (typically 0.4–1.0 μm), providing high spatial detail but no spectral information. Multispectral sensors (e.g., Landsat 8) capture 8–11 bands, each sensitive to a narrow wavelength range, enabling land-cover classification. Hyperspectral sensors (e.g., AVIRIS) record hundreds of narrow bands, creating a nearly continuous spectrum for each pixel—useful for mineral exploration, water-quality assessment, and precision agriculture. Higher spectral resolution enables finer discrimination between materials but increases data volume and processing complexity.

Radiometric resolution (or radiometric accuracy) refers to the number of discrete brightness levels a sensor can distinguish in each spectral band, expressed in bits. An 8-bit sensor can encode 256 levels (0–255); a 12-bit sensor can encode 4,096 levels. Higher radiometric resolution preserves subtle variations in reflectance (tone), improving detection of subtle surface changes and shadow detail. Modern sensors often use 12- or 14-bit radiometry. For example, an 8-bit Landsat band has coarser tone steps than a 12-bit WorldView band, potentially missing shadow boundaries in complex urban cadastre.

Temporal resolution is the time interval between successive observations of the same area (revisit time). Geostationary satellites (e.g., GOES, Himawari) revisit every 15–30 minutes, ideal for rapid-response disaster monitoring (typhoons, flooding). Polar-orbiting satellites (Landsat: 16 days; Sentinel-2: 5 days) provide moderate-frequency updates for crop monitoring and land-use change detection. The choice depends on the application: disaster response requires high temporal resolution; long-term environmental trend analysis may tolerate lower frequency. In the Philippines, typhoon season (June–November) drives high-temporal-resolution satellite procurement for flood and landslide forecasting.

Passive sensors detect naturally emitted or reflected electromagnetic energy and do not transmit their own signal. Optical sensors (visible and near-infrared) rely on sunlight reflection and work only during daylight with clear skies. Thermal infrared sensors detect heat emitted by objects and can operate day or night but are hampered by cloud cover. Examples include Landsat multispectral scanner, Sentinel-2 MSI (MultiSpectral Instrument), and MODIS on NASA satellites. Passive sensors are relatively inexpensive, have long operational histories (enabling multidecadal time-series analysis), and are excellent for vegetation and water mapping. However, they cannot penetrate clouds or operate at night.

Active sensors transmit their own electromagnetic energy and measure the returned signal (backscatter), independent of sunlight. Synthetic Aperture Radar (SAR) transmits microwave pulses and receives echoes, capable of penetrating clouds and vegetation and operating day/night. SAR is ideal for flood mapping under heavy cloud cover (common in Philippine typhoon season), subsurface soil-moisture assessment, and deformation monitoring via interferometry. Light Detection and Ranging (LiDAR) transmits laser pulses and measures round-trip time to create precise elevation models; it works through thin cloud and vegetation but is more expensive and operationally complex. Both provide geometric accuracy suitable for precise cadastral surveys and infrastructure management.

Raw satellite and aerial images contain geometric distortions due to sensor perspective, Earth curvature, terrain relief, and platform motion. Rectification corrects these distortions to align the image to a known map projection and datum (e.g., PRS92 with Philippine Transverse Mercator Projection—PPCS). This requires ground control points (GCPs): surveyed locations whose image and ground coordinates are known. Rectification transforms the image so that pixels align with geographic coordinates, enabling overlay with other spatial data and cadastral databases. A typical Philippine cadastral rectification might use 15–30 GCPs distributed across the project area to achieve sub-meter accuracy.

Enhancement improves the visual or analytical quality of an image through contrast stretching, histogram equalization, band combinations (e.g., false-color composite using near-infrared, red, green bands to highlight vegetation), and filtering. Classification assigns each pixel to a land-cover class (e.g., forest, urban, water, cropland) using spectral properties. Supervised classification uses training data (sample pixels of known class) and algorithms (maximum likelihood, support-vector machines) to label all pixels; unsupervised classification (k-means, ISODATA) groups similar pixels without prior training. Accuracy is validated against ground truth—GPS-surveyed or field-verified points. A typical cadastral classification in Philippine municipalities might identify parcel boundaries, building footprints, and land use with 85–95% accuracy.

The vector model represents geographic features as discrete objects: points (locations), lines (roads, boundaries, streams), and polygons (parcels, administrative units, water bodies). Each feature has geometry (shape and location) and associated attributes (name, owner, area, zoning class) stored in a database table. Vector data is ideal for cadastral databases, property boundaries, utility networks, and administrative divisions—features with well-defined, precise boundaries. For example, a Philippine cadastral parcel is represented as a polygon with vertices in PPCS coordinates, linked to attributes like lot number, owner name, and land classification (agricultural, residential, commercial) per RA 8560 (Residential Real Estate Installment Buyer Protection Act) or PD 1529 (Property Registration Decree) requirements. Vector's precision makes it suitable for legal and engineering applications.

The raster model represents the landscape as a regular grid of cells (pixels), each with a single value (elevation, temperature, vegetation index, classification code). Raster is ideal for continuous phenomena (elevation surfaces, temperature fields, rainfall interpolations) and remote-sensing imagery. Each cell's position is defined by row and column indices; the cell size determines spatial resolution. A 30 m resolution elevation raster for a municipality contains cells 30 m × 30 m, each holding a mean elevation value. Raster data is computationally efficient for overlays and statistical analysis but requires interpolation to estimate values between cells. A typical Philippine digital elevation model (DEM) for floodplain mapping might be a 10–30 m resolution raster created from LiDAR or SRTM data.

All spatial data must be referenced to a common datum and projection to ensure accurate overlay and analysis. The Philippines uses two primary systems: World Geodetic System 1984 (WGS84) for GPS and global data, and Philippine Reference System 1992 (PRS92) for official cadastral and engineering surveys. WGS84 is typically projected to Universal Transverse Mercator (UTM) zones; PRS92 uses the Philippine Transverse Mercator Projection (PPCS). GIS software includes transformation tools (Helmert or affine transformation using GCPs) to convert coordinates between systems. For example, a Landsat image (WGS84/UTM) must be transformed to PRS92 before overlaying a municipal cadastral database. Mismatched coordinate systems are a common source of GIS errors and a frequent PRC exam pitfall.

Overlay combines multiple vector layers to extract or create new spatial information. A union overlay merges all boundaries from two polygon layers; an intersection overlay retains only the area where both polygons overlap; a difference overlay removes one polygon layer from another. These Boolean operations are fundamental to land-suitability analysis, zoning compliance checking, and cadastral partition. For example, overlaying a flood-hazard polygon with a cadastral parcel layer identifies properties at risk. In the Philippines, overlaying zoning polygons (municipal land-use plan) with property boundaries (cadastral layer) identifies parcels requiring zoning-use variances.

Buffering creates a zone of specified distance around a feature. A 500 m buffer around a river polygon identifies all land within 500 m of the river (useful for riparian-zone protection under environmental laws). Buffer operations apply to points (e.g., a 100 m safety zone around a hazardous facility) and lines (e.g., a 50 m corridor along a road centerline for utility easements). Multiple buffers can be combined to create graduated impact zones. In the Philippines, buffering is used to define coastal-zone management areas under the Coastal Zone Management Act and to identify areas subject to building restrictions near fault lines (earthquake hazard zones).

Network analysis operates on connected linear features (roads, utility lines, streams) to solve problems such as shortest-path routing, facility location, and service-area delineation. Given a road network, GIS can identify the shortest route between two points, calculate travel times considering speed limits, and delineate areas serviceable by an emergency facility within a specified response time. In the Philippines, network analysis is used for emergency-response planning (ambulance routing), utility management (water-distribution optimization), and transportation planning (route optimization for jeepneys in Metro Manila).

Interpolation estimates values at unsampled locations from known sample points (e.g., surveyed elevation points, rainfall station data). Common methods include Inverse Distance Weighting (IDW), which weights nearby points more heavily; Kriging, a geostatistical method that models spatial autocorrelation; and Spline, which fits a smooth surface through points. The result is typically a continuous raster surface (DEM, rainfall map, contamination gradient). Interpolation quality depends on point density, spatial distribution, and method choice; areas far from sample points have lower confidence. A Philippine DEM created from surveyed benchmark elevations via kriging provides estimates of elevation and slope at all pixels, suitable for hydrologic and geotechnical analysis.

Spatial statistics quantify patterns and relationships in geographic data. Key concepts include spatial autocorrelation (the tendency of nearby values to be similar) measured by Moran's I or Geary's C; clustering indices (Getis-Ord G*) identifying hot spots; and spatial regression (e.g., geographic weighted regression) modeling how relationships vary across space. For example, property values in a metro Manila cadastral database often exhibit positive spatial autocorrelation—expensive properties tend to cluster in affluent neighborhoods. Spatial statistics inform zoning decisions, infrastructure siting, and equity analyses. Understanding spatial autocorrelation prevents spurious statistical conclusions (e.g., treating nearby parcel values as independent samples when they are strongly correlated).

Each material (vegetation, concrete, water) has a unique spectral signature across multiple wavelengths. Sensors with high spectral resolution (many narrow bands) can distinguish between spectrally similar classes (e.g., different crop types) more reliably than multispectral sensors. This relationship directly affects land-cover classification accuracy in cadastral and environmental applications.

Finer spatial resolution (smaller pixel size) enables detection of smaller features. A 30 m Landsat pixel cannot resolve individual building footprints, but a 1 m WorldView pixel can. Conversely, coarser resolution reduces data volume and processing time. Sensor selection balances these trade-offs based on project requirements.

High temporal resolution (frequent revisits) enables rapid detection of dynamic changes (flood extent, crop growth, vegetation loss). Low temporal resolution suits detecting long-term trends (urbanization, deforestation over years). The choice depends on the phenomenon's temporal scale.

Passive optical sensors fail in clouds and darkness; active radar/LiDAR sensors penetrate clouds and operate at night. In monsoon-prone Philippines, active sensors are preferred for critical applications (disaster response). Passive sensors are sufficient for clear-sky regions and long-term monitoring where temporal gaps are acceptable.

Raw satellite imagery contains geometric errors. Rectification using GCPs aligns it to PRS92 coordinates, enabling accurate overlay with cadastral, administrative, and utility vector layers. Poor rectification causes misalignment and invalid analysis results.

Vector overlay operations (intersection, union) are conceptually intuitive for legal/administrative applications (cadastral queries) but computationally intensive for large datasets. Raster overlays (map algebra) are fast and suit continuous analysis (risk mapping) but require rasterization of vector data, losing boundary precision. GIS often uses both models together—vector for cadastre, raster for analysis.

Supervised classification relies on training data (ground truth—field-surveyed points or high-resolution imagery). Limited or biased training data leads to poor classification accuracy. Extensive ground-truthing (especially in diverse terrain like Philippine archipelago) is essential and resource-intensive.

Layers in mismatched coordinate systems (e.g., Landsat in UTM, cadastre in PPCS) will not align correctly during overlay. Transformation to a common reference system (PRS92) is mandatory before any spatial analysis. This relationship is a frequent PRC exam trap.

IDW is simple but can produce artificial 'bullseyes' around sample points; Kriging incorporates spatial structure (semivariogram) and typically produces smoother surfaces; Spline fits a continuous function and is sensitive to outliers. Choice depends on data distribution and phenomenon (e.g., elevation surfaces are well-suited to Kriging; property values to geographic weighted regression).

Buffers around rivers, roads, or hazards define impact zones. Buffer widths are often specified by law (e.g., riparian buffers under environmental code). Incorrect buffer widths can lead to non-compliance and project delays.

Network analysis requires topologically correct vector data—connected line segments with proper node geometry. Topological errors (gaps, overshoots, undershoot in road network) produce incorrect routing results. Data validation is essential before network analysis.

Wider swaths (e.g., Landsat 185 km vs Spot 60 km) cover more ground per pass, reducing the number of scenes needed for area coverage and lowering project cost. Swath is a trade-off: wider swaths often mean coarser spatial resolution or higher cost.

High-resolution satellite imagery (QuickBird, Pleiades, 0.5–2 m) or aerial photography is rectified to PRS92 coordinates and used to identify building footprints, parcel boundaries, and land-use patterns. Vector polygons delineate each parcel, linked to a cadastral database containing owner names, lot numbers, and area (per RA 4374 requirements). Periodic updates using new imagery enable detection of illegal structures, subdivision, or boundary encroachments. This workflow is standard in Philippine municipal assessment offices and the Land Registration Authority (LRA).

Multispectral satellite imagery (Landsat 8, Sentinel-2, 10–30 m) is classified via spectral analysis to identify land covers (forest, grassland, water, urban, cropland) across an entire province or municipality. Vector polygons or raster grids represent each LULC class. Time-series analysis (comparing classifications from multiple years) reveals changes: deforestation, urban expansion, agricultural conversion. Results support municipal development planning, environmental monitoring, and carbon accounting (RA 9003, Solid Waste Management Act; RA 6969, Hazardous Substances and Hazardous Wastes Management Act).

Radar SAR imagery (Sentinel-1, ALOS PALSAR) penetrates clouds and detects water surfaces even under dense vegetation, enabling rapid flood-extent mapping during typhoons—critical in the Philippines' typhoon season. SAR data is orthorectified to PRS92, classified to delineate flood boundaries, and overlaid with cadastral parcels and building footprints to identify affected properties. Temporal SAR stacks enable flood-depth estimation via radar backscatter. Results guide evacuation planning, damage assessment, and post-disaster reconstruction prioritization.

LiDAR or SRTM/ASTER satellite elevation data is processed into a raster DEM (10–30 m resolution) that covers a watershed or municipality. Hydrologic analysis derives flow direction, flow accumulation, and drainage networks. Slope and aspect layers quantify terrain steepness and orientation (important for agricultural productivity, solar potential, and landslide risk). Floodplain delineation via hydrologic modeling identifies areas inundated at different flood frequencies (100-year flood, 500-year flood). This informs infrastructure design, building setbacks, and hazard zoning per Philippine building codes.

Time series of rectified satellite imagery (Landsat, Sentinel-2) spanning years or decades are compared to detect urban expansion, deforestation, agricultural abandonment, and infrastructure development. Change-detection algorithms (image differencing, classification change analysis) highlight pixels that have changed class between dates. Results reveal sprawl patterns, identify encroachments on protected forests or coastal zones, and support density and growth-boundary planning in metropolitan areas like Metro Manila and Cebu.

Multispectral satellite imagery is used to calculate vegetation indices (NDVI—Normalized Difference Vegetation Index) that quantify crop health and phenological stage across large agricultural areas. Time-series indices reveal crop growth patterns, detect pest/disease stress, and estimate yield potential. Results guide targeted irrigation, fertilizer application, and harvest timing—especially important in Philippine rice and sugarcane regions. GIS integration with farm-parcel data (vector polygons) enables field-level recommendations.

Vector road networks (centerlines, nodes) are created from aerial/satellite imagery or field survey and stored in GIS. Utilities (water pipes, electrical lines, telecommunications) are similarly vectorized. Network analysis calculates maintenance priorities based on asset age, condition, and criticality (distance from hospitals, schools). Buffering around utilities defines easement zones. Overlay with cadastral parcels identifies property-access issues and right-of-way conflicts. This workflow is standard for municipal public-works offices and utility companies in the Philippines.

Proposed development projects (subdivisions, industrial parks, mining operations) are assessed against environmental and zoning constraints. Vector layers (protected forests, wetlands, slope zones, administrative zoning) are overlaid with the proposed project site (vector polygon) to identify overlaps and conflicts. Buffer analysis identifies sensitive areas (schools, health centers, residences) within specified distances. Results support Environmental Impact Assessment (EIA) reports required under Philippine law (RA 6969, RA 9003) and inform zoning variance applications.

Synthetic Aperture Radar interferometry (InSAR) using repeat-pass SAR imagery (Sentinel-1, ALOS-2) detects millimeter-level ground deformation over time, revealing subsurface structural features, ground subsidence, and tectonic motion. Applications include monitoring land subsidence in agricultural areas, detecting volcanic inflation, mapping fault geometry, and detecting mining-induced deformation. Deformation maps (typically raster) are draped over DEM and imagery to visualize 3D structure and deformation dynamics.

Optical satellite time series (Landsat, Sentinel-2) distinguish forest from non-forest using spectral indices; SAR and LiDAR add structural information (biomass, canopy height). Deforestation is detected by comparing classifications across years. Results support forest inventory, conservation planning, and carbon accounting under international REDD+ (Reducing Emissions from Deforestation and forest Degradation) programs. Important for the Philippines' extensive forest estate and carbon-mitigation commitments.

High-resolution satellite imagery captures coastal bathymetry (in clear-water areas via spectral analysis) and shoreline position. Time series reveal accretion or erosion trends. Shoreline vectors enable proximity analysis to identify properties at risk from sea-level rise and coastal erosion (relevant to Philippine archipelago). Vector topographic/bathymetric data supports coastal hazard mapping (typhoon surge, tsunami), marine-resources management, and environmental-compliance monitoring per Coastal Zone Management Act.