Remote Sensing & GIS (CE-504 (B)) - Unit 1 Short Notes

UNIT 1: CORE CONCEPTS OF REMOTE SENSING & GIS

I. FUNDAMENTALS OF REMOTE SENSING

A. Electromagnetic Spectrum & Spectral Reflectance

• Electromagnetic (EM) Spectrum: Range of EM radiation from gamma rays to radio waves. Remote sensing primarily uses visible, infrared (IR), and microwave regions.

• Spectral Reflectance (ρ): Ratio of reflected radiation to incident radiation on a surface. Varies with wavelength, creating a unique spectral signature for each material.

• Typical Spectral Reflectance Curves:

  • Soil: Generally low, increasing with wavelength. Influenced by moisture, texture, organic matter, and iron oxide content.

  • Vegetation: Low in blue/red (chlorophyll absorption), high in green (peak), very high in near-infrared (NIR) due to cell structure. Red-edge transition is critical.

  • Water: Very low in NIR and SWIR (strong absorption). Increases slightly in blue/green but decreases with turbidity/depth.

  [!TIP] Exam Focus: Be ready to sketch generalized curves for all three on the same graph, labeling key absorption/reflection features (e.g., red edge for vegetation).

B. Ideal Remote Sensing System

A conceptual system for efficient data acquisition and information extraction. Functional Components:

1. Energy Source: Sun (passive) or onboard transmitter (active).

2. Atmosphere-Earth Surface Interaction: Radiation modifies via absorption, scattering.

3. Sensor: Detects and records reflected/emitted EM energy.

4. Data Processing & Correction: Radiometric & geometric correction to generate usable imagery.

5. Information Extraction & Analysis: Visual/digital interpretation to derive information.

6. Application: Utilized for specific user needs (e.g., mapping, monitoring).

[!TIP] Common Pitfall: Students often forget the atmosphere as an active component modifying the signal before it reaches the sensor.

C. Sensor Resolutions

Four fundamental resolutions define a sensor's capability. Trade-offs exist; improving one often degrades another.

[!TIP] Exam Tip: Remember "SpaTial" for Spatial, Temporal. "SpecTral" for Spectral. "RadiomeTric" for Radiometric.

II. REMOTE SENSING PLATFORMS & SENSORS

A. Satellite Orbits & Characteristics

Synoptivity: Ability to image a large area in a single acquisition.

Repetivity: Frequency of revisiting the same area.

B. Indian Satellite Programs (IRS & Chandrayaan-3)

• IRS (Indian Remote Sensing) Series:

  • Objectives: Resource management (agriculture, forestry, water), environment monitoring, urban planning, disaster management.

  • Features: Variety of sensors (LISS, AWiFS, WiFS, HySIS). Sun-synchronous orbits. Progressive improvement in spatial (from 72m to <5m), spectral, and temporal resolutions.

• Chandrayaan-3 (Lunar Mission):

  • Objectives: Safe and soft landing on Moon's south polar region, rover operations, in-situ scientific experiments.

  • Salient Features: Lander (Vikram), Rover (Pragyan), propulsion module. First mission to land near lunar south pole. Studied lunar surface composition, exosphere.

C. Earth Observation Satellites

III. IMAGE INTERPRETATION & CLASSIFICATION

A. Visual Interpretation

Elements of Interpretation:

1. Tone/Color: Relative brightness/color in B&W or color composites.

2. Texture: Roughness/smoothness (e.g., forest = coarse, water = smooth).

3. Shape: Geometric form (e.g., buildings = rectangular, rivers = sinuous).

4. Size: Relative/absolute dimensions.

5. Pattern: Spatial arrangement (e.g., orchards = regular, natural forest = irregular).

6. Association: Relationship with other features (e.g., airport with runways, terminals, roads). Tools:

• Stereoscope: For 3D viewing from stereo pairs (parallax for height/depth).

• Color Composites: e.g., False Color Composite (FCC) using NIR, Red, Green bands (NIR=Red, Red=Green, Green=Blue). Vegetation appears red, water dark blue/black.

B. Digital Image Analysis

C. Classification Methods

• Supervised Classification:

  • Process: User selects training samples (representative areas of known land cover). Algorithm (e.g., Maximum Likelihood) learns spectral signature and classifies entire image.

  • Steps: 1. Define classes, 2. Collect training data, 3. Train classifier, 4. Classify image, 5. Evaluate accuracy (confusion matrix).

  • Use: When prior knowledge of classes exists.

• Unsupervised Classification (Clustering):

  • Process: Algorithm (e.g., ISODATA, K-means) groups pixels into clusters based on spectral similarity without training data. User then interprets/clusters into information classes.

  • Steps: 1. Specify number of clusters, 2. Run clustering, 3. Analyze cluster properties, 4. Merge/assign to meaningful classes.

  • Use: Exploratory analysis, unknown area, generating spectral classes.

[!TIP] Key Difference: Supervised uses training data to define classes; Unsupervised lets the data define clusters.

IV. GIS FUNDAMENTALS

A. Definition, Objectives & Components

• Definition: GIS is a computer-based system for capturing, storing, managing, analyzing, and displaying geographically referenced data.

• Objectives: Support decision-making for spatial problems, integrate diverse data, automate mapping, perform spatial analysis, model scenarios.

• Five Core Components:

  1. Hardware: Computer, storage, GPS, plotters.

  2. Software: GIS application (ArcGIS, QGIS), DBMS, tools.

  3. Data: Spatial data (features with location) + Attribute data (descriptive info). Most critical and expensive component.

  4. People: skilled users, managers, decision-makers.

  5. Procedures: Workflow, data standards, quality control.

B. Data Sources & Collection Objectives

Primary Objective of Data Collection: To obtain accurate, relevant, and current spatial and attribute data that meets the specific requirements of the GIS application.

C. Spatial & Attribute Data Integration

• Integration Mechanism: Each spatial feature (point, line, polygon) is assigned a unique identifier (Feature ID or Primary Key).

• Linkage: The attribute table contains a Foreign Key (same as Feature ID) that links each record to its corresponding spatial feature.

• Relational Database Management System (RDBMS): Manages attribute tables. Allows queries like "Show all wells (spatial) where yield > 100 L/min (attribute)" by joining tables on the common key.

[!TIP] Conceptual Link: Spatial Data answers "WHERE?" and "WHAT SHAPE?". Attribute Data answers "WHAT?", "WHEN?", "HOW MUCH?".

V. GIS DATA MODELS & STRUCTURES

A. Vector vs. Raster Data Structures (Detailed Comparison)

B. Data Models in GIS

• Feature-Based (Vector) Model: Represents real-world entities as discrete objects with geometry (coordinates) and attributes.

• Grid-Based (Raster) Model: Represents continuous surfaces or phenomena as a matrix of cells, each with a value representing a property (e.g., elevation, land cover).

C. Data Conversion in GIS

• Vector-to-Raster (Rasterization): Convert vector features to grid cells.

  • Method: For each cell, determine which feature it falls within (point-in-polygon). Assign cell value from feature's attribute.

  • Use: Preparing vector data for raster-based analysis (e.g., suitability modeling).

• Raster-to-Vector (Vectorization): Convert raster grid to vector features.

  • Method: Edge detection to find boundaries, then line following to create arcs, polygonization to close loops.

  • Use: Creating editable vector maps from scanned images or classified imagery.

D. Shapefile Creation Procedure

1. Digitization:

   • Manual: Using a digitizing tablet to trace features from a hardcopy map.

   • On-Screen (Heads-up): Tracing features directly from a georeferenced raster image (e.g., satellite image, scanned map) in GIS software.

2. Attribute Assignment: During/after digitization, populate the attribute table. Each feature gets a unique ID. Add descriptive fields (e.g., Road_Name, LandUse_Type).

3. Topology Building (Optional but Recommended): Define rules (e.g., no dangles, no overlaps) and build topology to ensure data integrity for network or polygon analysis. Creates topological relationships (arcs, nodes, polygons).

4. Saving as Shapefile: The final vector layer is saved as a .shp (geometry), .shx (index), .dbf (attributes), and .prj (projection) file set.

VI. SPATIAL REFERENCE & ANALYSIS

A. Coordinate Systems

B. Map Projections

• Importance: Earth is 3D, maps are 2D. All projections introduce distortion in shape, area, distance, or direction. Choosing a projection depends on the map's purpose (e.g., conformal for navigation, equal-area for thematic maps).

• UTM Projection:

  • System: Global system of 60 zones (6° wide, numbered 1-60 from 180°W).

  • Characteristics: Transverse Mercator projection. Conformal (preserves shape locally). Scale factor of 0.9996 at central meridian to minimize distortion. Each zone has its own central meridian and false easting/northing (in meters) to avoid negative coordinates.

  [!TIP] UTM Zone Calculation: Zone number = floor((Longitude + 180) / 6) + 1. For India (68°E to 97°E), Zones 43 to 47.

C. Spatial Analysis Techniques

• Buffer Analysis:

  • Concept: Create a zone of specified distance around a feature (point, line, polygon).

  • Applications: Riparian buffer (100m from river), noise pollution zone (500m from highway), service area of a hospital (5km drive time).

• Overlay Analysis:

  • Concept: Combine two or more thematic layers to identify relationships. Operates on overlapping polygons.

  • Types:

    1. Intersection: Output retains only the area where all input layers overlap. Features get attributes from all layers. (Most common).

    2. Union: Output retains all areas from all input layers. Features get attributes from the layer they came from.

    3. Identity: Like intersection, but retains all features from the identity layer (first input), with attributes from the second layer where they overlap.

  [!TIP] Mnemonic: Intersection = In common. Union = Unites everything.

D. Errors in GIS

[!TIP] Key Difference: Systematic error has a pattern (can be corrected); non-systematic is random (reduced by better procedures).

VII. APPLICATIONS & INTEGRATION CHALLENGES

A. Land Use/Land Cover (LULC) Change Assessment (Integrated RS-GIS Workflow)

1. Data Acquisition: Obtain multi-temporal satellite images (e.g., Landsat 2000, 2020) for the study area.

2. Pre-processing: Perform atmospheric correction and geometric correction/orthorectification (using DEM) to align images perfectly.

3. Image Classification: Apply supervised or unsupervised classification separately for each date to generate LULC maps (e.g., classes: agriculture, urban, forest, water).

4. Post-classification Processing: Clean classified maps (remove speckle, generalize).

5. GIS Overlay & Change Detection: In GIS, perform overlay analysis (intersection) of the two classified raster/vector layers. Create a change matrix (cross-tabulation) to quantify gains/losses for each class.

6. Analysis & Mapping: Identify change hotspots, calculate rates of change, map conversion types (e.g., forest → agriculture).

B. Water Resources Applications

• Watershed Delineation: Use Digital Elevation Model (DEM) from RS (e.g., SRTM, ASTER) in GIS to derive flow direction, accumulation, and automatically delineate watershed boundaries and stream networks.

• Groundwater Mapping: Identify potential recharge zones (lineaments, fractures from satellite imagery), map aquifer extent (using geophysical data integrated in GIS), and model groundwater flow.

• Flood Assessment & Management:

  • Inundation Mapping: Use SAR (Sentinel-1) or optical imagery during flood to map extent.

  • Risk Zoning: Combine flood hazard maps (from hydraulic models) with exposure data (population, infrastructure) in GIS.

  • Floodplain Delineation: Use DEM and hydraulic models in GIS.

C. Traffic Management Applications

• Route Optimization: GIS network analysis (shortest path, vehicle routing problem) for emergency services, delivery fleets.

• Congestion Monitoring: Use high-temporal resolution RS (e.g., from traffic cameras, or SAR for large-scale flow) to estimate traffic density/speed. Integrate with GIS to visualize congestion hotspots.

• Infrastructure Planning: Optimal location for new roads, intersections, or public transport hubs using suitability modeling (overlay of slope, land use, existing network).

D. Problems in Using Remote Sensing Data in GIS

[!TIP] Golden Rule: "Garbage In, Garbage Out (GIGO)." The quality of GIS analysis is fundamentally limited by the quality and compatibility of the input RS data. Always pre-process and validate RS data before integration.