Multispectral Remote Sensing Systems

Learning Materials and Assignment Overview

Prescribed Reading: This study unit (Study Unit 9) corresponds to Chapter 7 of the prescribed textbook. Students are encouraged to obtain a copy of this chapter from the library for additional information.
Practical Context: The concepts discussed relate directly to the data used in practical exercises, including data from Landsat, SPOT, and WorldView systems.
Assignment Details: A theoretical and research-based assignment is linked to this unit. Its requirements include: - Application Definition: Choose an application area such as agriculture or meteorology. - Problem Statement: Define a specific problem to address using multispectral remote sensing data. - Sensor Selection: Select at least two remote sensing systems of interest. - Parameter Research: Research and summarize parameters for these sensors (e.g., from specification documents or manuals found via Google). - Calculations: Perform specific calculations based on the chosen sensors' parameters.

Definition and Characteristics of Multispectral Remote Sensing

Formal Definition (Jensen): Multispectral remote sensing refers to a system that records reflected or emitted radiant energy from an object or area of interest in multiple spectral bands.
Spectral Characteristics: - Band Count: These systems typically record in fewer bands, ranging from 2 to 30+ discrete bands (e.g., Landsat or Sentinel-2). - Bandwidth: Multispectral sensors utilize broad bandwidths, often reaching up to $100\,nm$ . For example, the SPOT green band ranges from $0.5$ to $0.6\,\mu m$ , which translates to a width of $100\,nm$ .
Limitations in Discrimination: Due to broad bandwidths, multispectral systems do not provide precise enough radiometric information to distinguish between virtually similar targets, such as different tree species or specific soil types (e.g., clay vs. sandy soil). However, they are ideal for mapping broad land cover classes like vegetation, water, and soil.

Multispectral vs. Hyperspectral Systems

Hyperspectral Systems: These systems, such as the Hyperion sensor, record data in many narrow, contiguous bands, typically ranging from 30 to 300+ bands (Hyperion has approximately 220 bands).
Spectral Resolution: Hyperspectral sensors have higher spectral resolution due to narrower bandwidths (potentially as narrow as $1\,nm$ ), allowing for the discrimination of targets based on subtle biological or chemical properties like chlorophyll or water content.
Spectral Signatures: Multispectral sensors produce "rough" or non-contiguous spectral curves, whereas hyperspectral sensors produce smooth, contiguous spectral signatures that show fine separability between targets.

Key Components and the Scanning Process

Illumination Source: The Sun provides the necessary radiant energy for the system.
Spacecraft and Sensor: The sensor is mounted on a platform (satellite, drone, or aircraft). A key component is the receiver, which is a mirror that captures reflected radiance from the ground.
Scanning Motion: The mirror oscillates perpendicular to the flight direction (across-track) to scan the terrain.
Detectors and Optics: Reflected radiance is channeled through optics to an array of detectors. These detectors split the light into specific spectral components (green, red, infrared, shortwave infrared) and store the energy data.
Data Transmission and Conversion: Data is initially captured in analog form. During a satellite overpass, ground station receivers (satellite dishes) lock onto the satellite and download the data. This data is then converted from analog to digital format, resulting in the Digital Numbers ( $DN$ ) or Brightness Values used in analysis.

The South African National Space Agency (SANSA)

Space Operations: SANSA (formerly the Satellite Application Center) at Hartebeeshoek operates receivers to track satellites and download data for South Africa.
Data Storage: Data is stored on large servers and can be requested at various processing levels, including: - Raw data. - Radiometrically corrected data. - Geometrically corrected data. - Reflectance data.

Scanner Mechanics: Whiskbroom vs. Pushbroom Systems

Whiskbroom (Across-track) Scanners: - Use a rotating/oscillating mirror to scan ground strips perpendicular to the flight direction. - Examples: MODIS, AVHRR, and older Landsat missions. - Advantages: Can cover very wide areas (swaths). - Disadvantages: Contains moving parts that can deteriorate; inherent geometric distortion (pixel elongation) occurs at the edges of the swath due to the wide total field of view.
Pushbroom (Along-track) Scanners: - Consists of a fixed linear array of detectors that records an entire line at a single instant as the satellite moves forward. - Advantages: No moving parts; allows for a longer "dwell time," which improves the signal-to-noise ratio and permits higher spatial resolution and narrower bandwidths. - Disadvantages: Generally covers a narrower swath compared to whiskbroom systems.

Calculating Swath and Spatial Parameters

Total Field of View ( $TFOV$ ): This is the angular extent (measured in degrees) of the scanner. It determines the swath width (the width of the ground area scanned).
Swath Example (Landsat): - $TFOV = 15^{\circ}$ - Flying height ( $H$ ) = $705\,km$ - Resulting Swath width = $185\,km$
Instantaneous Field of View ( $IFOV$ ): The angular cone of visibility of a single detector, usually measured in radians ( $\beta$ ). This defines the ground projected area ( $d$ ) or the size of a single pixel.
Formula for Ground Resolution Cell ( $d$ ): - $d = \beta \times H$

Understanding Dwell Time

Definition: Dwell time is the amount of time a sensor spends observing and recording radiant energy from a single ground resolution cell.
Formula for Dwell Time: - $\text{Dwell Time} = \frac{\text{Down-track pixel size}}{\text{Orbital velocity}} \times \frac{\text{Cross-track pixel size}}{\text{Cross-track line width (Swath)}}$
Calculation Constraints: All units must be consistent (e.g., convert kilometers to meters). For Landsat, the down-track and cross-track pixel sizes are typically $30\,m$ , and the swath is $185,000\,m$ .
Significance: Increased dwell time allows more energy to be detected, improving radiometric quality.

Satellite Orbits: Geostationary and Sun-Synchronous

Geostationary Orbits: - Position: Satellites are locked over a specific portion of the Earth, rotating at the same speed as the planet. - Altitude: Approximately $36,000\,km$ . - Applications: Meteorology (e.g., Meteosat for Africa/Europe, GOES for the Americas) and telecommunications (e.g., DSTV signals). - Resolution: Very high temporal resolution (images every 5 to 15 minutes) but coarse spatial resolution (often $1\,km$ or more), making it ideal for tracking rapidly changing phenomena like clouds.
Sun-Synchronous (Polar) Orbits: - Position: Satellites orbit from North to South, synchronized to pass over the Earth at the same local solar time. - Altitude: Ranges from $300\,km$ to $1,000\,km$ (e.g., Landsat at approximately $705\,km$ ). - Applications: Environmental monitoring, land cover change, and natural resource management. - Resolution: High spatial and spectral resolution compared to geostationary satellites, but lower temporal resolution (revisit times are measured in days or weeks).

Orbital Dynamics and Variations

Orbital Tracks: Satellites move along specific orbital tracks; for Landsat, these are cataloged via the World Reference System ( $WRS$ ), which allows users to identify specific "scenes" globally.
Apogee and Perigee: An orbit has a point where the satellite is farthest from the Earth (apogee) and closest (perigee). These distance variations cause illumination changes, which are corrected using a systematic de-correction factor at ground stations.
Morning vs. Afternoon Orbits: Some satellites operate in the morning, others in the afternoon. For example, the MODIS sensor is mounted on two satellites: Terra (morning orbit) and Aqua (afternoon orbit), providing two terrestrial views per day.