Remote sensing is the science and art of obtaining qualitative and quantitative information about the Earth's surface without entering into direct physical contact with the target. By deploying sensors aboard satellites, aircraft, and unmanned aerial vehicles (UAVs), remote sensing records the electromagnetic radiation (EMR) reflected, emitted, or backscattered from Earth features. Because different materials—such as green healthy vegetation, clear water, turbid sediment, dry soil, and asphalt—interact with electromagnetic wavelengths in uniquely characteristic ways, we can identify and map land cover dynamics globally. This lecture breaks down the electromagnetic spectrum, atmospheric interaction, and sensor resolution dimensions.
📋 Prerequisites
- Basic understanding of physics and light waves.
- Familiarity with digital imagery and pixel concepts.
- Interest in Earth observation satellite systems.
🛠️ Technical Environment
Required Software: Python / Rasterio & QGIS (Recommended: Python 3.10+ / QGIS 3.34 LTR)
Practice Dataset: Copernicus Sentinel-2 Level-2A Surface Reflectance Tile
Source Portal: ESA Copernicus Data Space Ecosystem
CRS / Format: UTM Zone (Local) (Cloud-Optimized GeoTIFF (COG))
Step-by-Step Workflow & Methodological Execution
Module 1: The Electromagnetic Spectrum & Atmospheric Windows
Electromagnetic radiation travels through space as sinusoidal waves characterized by wavelength (λ) and frequency (ν). Remote sensing utilizes specific bands of the EM spectrum: 1. Visible Light (0.4 to 0.7 µm): Divided into Blue (0.4-0.5 µm), Green (0.5-0.6 µm), and Red (0.6-0.7 µm). This is the only portion detectable by the human eye. 2. Near-Infrared / NIR (0.7 to 1.1 µm): Crucial for vegetation monitoring due to high structural scattering in leaf spongy mesophyll cells. 3. Shortwave Infrared / SWIR (1.1 to 3.0 µm): Highly sensitive to moisture content in vegetation and soil, as well as mineral differentiation. 4. Thermal Infrared / TIR (3.0 to 14.0 µm): Measures emitted radiant kinetic temperature rather than reflected sunlight, used for urban heat islands and sea surface temperatures. 5. Microwave (1 mm to 1 m): Utilized by Synthetic Aperture Radar (SAR) systems to penetrate clouds, smoke, and nighttime conditions. Importantly, Earth's atmosphere absorbs specific wavelengths due to ozone, water vapor, and carbon dioxide. Sensors must be calibrated to operate within 'Atmospheric Windows'—wavelength ranges where the atmosphere is highly transparent.
Module 2: Spectral Reflectance Curves (Signatures)
A spectral reflectance curve plots the percentage of incident electromagnetic radiation reflected by a material across various wavelengths: • Green Vegetation: Characterized by strong absorption in the Blue and Red bands by photosynthetic chlorophyll pigments, a moderate peak in Green (~550 nm), a sharp rise known as the 'Red Edge' (~700 nm), and an immense plateau of high reflectance in the NIR (700-1100 nm) caused by internal leaf structure scattering. • Clear Water: Absorbs nearly all incident radiation in the NIR and SWIR regions. Water appears very dark or black in infrared imagery, allowing sharp, automated land-water boundary delineation. • Bare Soil: Displays a steady, monotonic increase in reflectance from visible to SWIR wavelengths, interrupted by diagnostic water absorption troughs at 1.4 µm and 1.9 µm.
Module 3: The Four Dimensions of Sensor Resolution
Every Earth observation satellite system is defined by four core resolutions that dictate its operational utility: 1. Spatial Resolution: The physical ground dimension represented by a single pixel (Ground Sample Distance - GSD). Sentinel-2 provides 10m visible bands; Landsat 8 provides 30m; commercial sensors like WorldView-3 offer 0.3m. 2. Spectral Resolution: The number, bandwidth, and placement of electromagnetic bands captured. Multispectral sensors capture 4 to 13 discrete wide bands; Hyperspectral sensors capture hundreds of contiguous narrow bands (5-10nm wide). 3. Radiometric Resolution: The sensor's sensitivity to subtle differences in radiant energy, measured in bits. 8-bit imagery yields 256 gray levels; modern 12-bit and 16-bit sensors record 4,096 to 65,536 discrete intensity levels, preserving details in shadows and bright clouds. 4. Temporal Resolution: The revisit time required for a satellite to image the exact same geographic location on Earth (e.g., Landsat is 16 days; Sentinel-2 constellation is 5 days; PlanetScope is daily).
⚠️ Common Errors & Troubleshooting
❌ ZeroDivisionError in NIR/Red ratio
💡 Resolution: Wrap ratio operations in numpy with 'np.errstate(divide='ignore', invalid='ignore')' and mask zero-sum pixels.
❌ Radiometric values exceed 1.0
💡 Resolution: Sentinel-2 L2A BOA reflectance values are stored as integers with a 10,000 scale factor; divide raw DN by 10,000 to obtain true surface reflectance.
💡 Expert Tips & Best Practices
- Use Band 4 (Red) and Band 8 (NIR) for 10-meter resolution spatial analysis.
- Check scene cloud cover metadata (tile metadata XML) before running automated land surface classification.
🐍 Python Rasterio Band Ratio & Reflectance Extraction
import rasterio
import numpy as np
# Open a multispectral satellite GeoTIFF (Sentinel-2 L2A)
with rasterio.open("sentinel2_scene.tif") as src:
# Read Red (Band 4) and Near-Infrared (Band 8)
red = src.read(4).astype(float)
nir = src.read(8).astype(float)
profile = src.profile
# Prevent division by zero and calculate simple NIR/Red ratio
with np.errstate(divide='ignore', invalid='ignore'):
ratio = np.where((nir + red) == 0, 0, nir / red)
print(f"Scene Dimensions: {src.width} x {src.height} pixels")
print(f"Coordinate Reference System: {src.crs}")
print(f"Mean NIR/Red Ratio over scene: {np.nanmean(ratio):.2f}")