Skip to main content
ARS Home » Northeast Area » Beltsville, Maryland (BARC) » Beltsville Agricultural Research Center » Hydrology and Remote Sensing Laboratory » Docs » RSBasics » Research » Page 3

Page 3
headline bar
1 - Research
2 - Page 3
3 - Page 4
4 - Page 5
5 - Page 6
6 - Page 7
7 - Page 8
8 - Page 2
9 - Page 9
10 - Page 10
Remote Sensing Basics

I. Sources of Imagery

II. Signal-to-Noise Ratio (S/N)

III. Imaging Theory: What to Expect

IV. Simple Tools for Checking Image Data Quality

V. Choice of Spectral Bands

VI. Final Thoughts and Suggestions

VII. Resources and References




II. A Working Philosophy for Airborne Remote Sensing: S/N

Working with electro-optical airborne remote sensing is fundamentally about Signal-to-Noise ratio (S/N): how much of the recorded signal that appears as a pixel is useable information, and how much is unwanted distortion or noise. Understanding the remote sensing process, limitations of technology, and knowledge of the target provide guidance for this approach.

The Remote Sensing

A review of the remote sensing process with a S/N perspective is a good place to start. The remote sensing process consists of:

  1. the sun as a source of radiant energy,
  2. transmission of solar radiation through the atmosphere,
  3. interaction of the solar radiation with the surface,
  4. transmission of reflected solar radiation back through the atmosphere towards the sensor,
  5. interception of the radiation by the sensor, and
  6. analysis.

The process is a system with input (solar energy) and output (information). Each component of the system modifies or adds to the signal. Thus, each component has its own modulation transfer function or MTF. Although MTF is most often associated with image/signal processing theory, the idea that the atmosphere, earth's surface, and the electronics modulate the information content of the energy flow through the system is the same. We can thus speak of an atmospheric MTF, a surface MTF, and an instrumentation system MTF. The modulation process induces a noise component to the signal that is sensed by the detector and made available to the analyst.

The quantity of radiant solar energy received by the earth is very stable over any given six-month period. Our concern with the sun from a S/N perspective concentrates on the effects of solar zenith and azimuth changes on the direction and size of shadows, and the amount of energy that reaches a particular surface. This is why airborne remote sensing data collection (including aerial photography) at optical wavelengths is usually limited to plus or minus 2 hours from solar noon. The geometry of illumination relative to viewing geometry modulates the signal. These issues fall under the realm of the bidirectional reflectance distribution function (BRDF).

The atmosphere modulates the incoming or downwelling radiant solar energy via scattering and absorption before the energy interacts with the earth surface. More haze translates to more diffuse radiation than direct beam solar radiation reaching the surface. From the S/N perspective, this means that shadows not be as dark.

The interaction of solar energy with earth surface materials can be viewed as a modulation. Multiple factors contribute to the modulation, especially for the case of plant canopies. One way to understand what factors are important to energy interactions is to examine inputs to radiative transfer (RT) models used to calculate reflectance from surfaces. Plant canopy RT models such as the SAIL model are especially helpful. Inputs include the foliage density expressed as leaf area index (LAI), leaf reflectance and transmittance, soil reflectance, information about leaf angles relative to vertical, and the amount of diffuse solar radiation relative to direct beam solar radiation (as per the atmospheric MTF above). This explains why spectral vegetation indices are sensitive to a variety of characteristics such as plant foliage density, leaf reflectance, percent vegetative ground cover, and soil brightness.

Solar energy is reflected back into the atmosphere following interaction with surface materials Once again, there is modulation of the signal as the energy is scattered and absorbed by atmospheric constituents. The result of the atmospheric modulation is the addition of the atmospheric signal as haze to the information from the surface interactions.

The sensor system with its optics, detectors, spectral discrimination system and electronics is a a form of transducer. A transducer converts one form of energy into another. Sensor systems convert radiant energy into electrical energy. Thus, the output of the sensor system is a voltage. The voltage is a function of all of the prior remote sensing system components and the electro-optical sensor system. The optical subsystem and the electrical subsystem both add noise to the sensing process, thus further modulating the information signal from the environment.

Note that the sensor platform (fixed wing aircraft or helicopter), also induces noise via attitudinal changes. Roll, pitch and yaw changes while scanning induce geometric distortions, and if extreme enough, BRDF noise.

The analysis of sensor output can also be thought of as an opportunity for signal modulation. Enhancements, data compression, transformations, etc. can induce noise that hinders the extraction of desired information.

The modulation of the energy by the sun, atmospheric, target, and electro-optical system can be thought of as the process by which noise is induced to the final output of the sensor that undergoes analysis. A radiometric measurement expressed as a pixel can be though of as a sum of the desired information signal and a noise signal.

So in effect, each step or element of the remote sensing process introduces noise to the signal. A common goal of electro-optical system engineering, data collection, system calibration, and data processing is to maintain employ techniques that produce a high fidelity system: tools that reproduce the input information faithfully, or accurately.

WebMaster   |  Content Manager

<< Previous 1 [2] 3 4 5 6 7 8 9 10 Next >>