Introduction to Remote Sensing
Student Learning Outcomes Define remote sensing. Outline the major steps in the remote sensing process in correct chronological order. List some ways in which remote sensing can inform field work and vice versa. Discuss some of the major strengths and limitations of remote sensing. Discuss some of the major strengths and limitations of field (in situ) approaches.
Student Learning Outcomes Differentiate between biophysical and hybrid remote sensing variables. Contrast active and passive remote sensing systems. Define spatial resolution. Define spectral resolution. Define temporal resolution. Define radiometric resolution. Specify the remote sensing data characteristics most desirable for a given set of research objectives (be sure to acknowledge trade- offs in terms of the four kinds of resolution, availability, cost, and swath). Explain your choices.
Definition of Remote Sensing … art, science, and technology of obtaining information about the environment through the recording, analysis, and interpretation of digital representations of energy patterns derived from noncontact sensors ~ Colwell 1997 Remote sensing instrument = sensor
Remote Sensing: A Science Science ~ Field of human knowledge Hypotheses, laws, theories, scientific method, … At the interface of natural and social sciences Physical Sciences Sciences Physical, biological, social sciences Mathematics and logics = sciences or tools for science? Remote sensing Tool, Technique, Scientific activity Biological Sciences Jensen 2005: 4.
Remote Sensing: A Science Developmental Stages of a Scientific Discipline Sigmoid / logistic curve describes Remote Sensing Growth rate of publications * Adapted from Jensen 2000: 5. Little / no social organization Growth of collaboration Increasing specialization + controversy Membership decline Developmental Stages of Scholarly Field
Remote Sensing: An Art Quality of image analyst Understanding scientific principles Real-world knowledge Synthesis of scientific principles & Real-world knowledge Logical conclusions
A Remote Sensing Scientist Intelligence Dweeb Social ineptitude Nerd Geek Dork Obsession
The RS Process in a Nutshell Observe a phenomenon State the problem State the objectives (Formulate hypothesis) State the significance Design the methodology Collect the data Analyze the data (Test hypothesis) Interpret the results Present the results Discuss the results Conclude your work
The RS Process in a Nutshell
Remote Sensing: Pros & Cons Advantages Unobtrusive (passive sensors) Systematic data collection – less bias than in some in situ studies Provides spatially continuous data Synoptic view – may provide data for large and often remote areas Data collection across EM spectrum (visible and more) Historical perspective Limitations Intrusive (active sensors) Human method-induced errors (e.g., in sensor selection, calibration) Cost of data and data collection, analysis, and interpretation No magic bullet … and what about in situ (i.e., field) approaches? Pros & cons Relationship with remote sensing approaches
Data Requirements In situ data Ancillary / collateral data Collected in field and/or lab Used for calibration Used for evaluation Ancillary / collateral data Often used to aid with calibration Remote sensing data Used to map biophysical variables: measured directly (e.g., land surface temperature, leaf area index, or dissolved organic matter) Used to map hybrid variables: measured indirectly (e.g., land cover, land use, or vegetation stress) Collected with a diversity of remote sensing systems
Many Types of RS Systems Analog Digital Ana
Active Remote Sensing Microwave energy is transmitted toward an area from an antenna in very short bursts or pulses Energy is reflected off of the area (backscattered) and its strength (detection) recorded at the antenna after a certain time delay (ranging) Propagation of one radar pulse
Passive Remote Sensing Space Sensor Sun Atmosphere Earth’s Surface
Spectral Reflectance 16
Spectral Reflectance 17
Remote Sensing Data Collection Passive sensors Detect electromagnetic radiation that is naturally reflected (visible, near-infrared, shortwave infrared) or emitted (thermal infrared) by objects Energy source = sun Active sensors Detect electromagnetic radiation that is reflected (i.e., backscattered) from objects that are irradiated from artificially generated energy sources Energy sources = Radio, Sound, or Light Detection and Ranging systems
Which Data to Acquire? Depends on research objectives Four types of resolution Spatial Resolution Spectral Resolution Temporal Resolution Radiometric Resolution Other Availability for place and time Cost Swath width ...
Spatial Resolution Measure of the degree of detail in a spatial data set Various definitions: size of smallest detectable feature; one half the size of the smallest detectable feature; minimum angular or linear distance between detectable features ~ Pixel size (smallest non-divisible element of digital image) Columns (j) 82 30 m 3 40 53 80 5 2 35 50 82 15 17 25 13 18 14 Rows (i)
Spatial Resolution 1 × 1 m
Spatial Resolution 2 × 2 m
Spatial Resolution 3 × 3 m
Spatial Resolution 5 × 5 m
Spatial Resolution 10 × 10 m
Spatial Resolution 15 × 15 m
Spatial Resolution 30 × 30 m
Spatial Resolution 1 × 1 m 3 × 3 m 5 × 5 m 10 × 10 m 15 × 15 m
Spatial Resolution Rule of thumb for feature detection: Spatial resolution < ½ size of smallest feature to be detected Tree crown diameter = 4 m Spatial resolution < 2 m
Matrix (Array) of Digital Numbers (DNs) Spectral Resolution Number and dimension (size) of specific wavelength intervals (i.e., bands or channels) in the EMS to which a remote sensing instrument is sensitive Matrix (Array) of Digital Numbers (DNs) Columns / Samples (j) 1 10 15 17 20 21 15 16 18 21 23 2 17 18 20 22 22 3 18 20 22 24 25 Rows / Lines (i) 4 y-axis (i rows) x-axis (j columns) Bands (k) (co-registered)
Spectral Resolution × “Spectral signatures” ×
Spectral Resolution Spectral Resampling X
Spectral Resolution Multi-, Hyper-, and Ultra-Spectral RS Systems Collect reflected or emitted energy from features or areas of interest, typically in digital format Multispectral: Multiple (a few; > 2) wide, separated bands Hyperspectral: Hundreds of fairly narrow, contiguous bands Ultraspectral: Thousands of very narrow, contiguous bands Major difference Not so much the number of measured wavelengths but the narrowness and contiguous nature of the measurements
Temporal Resolution Return interval of a remote sensing system to a particular location June 1, 2010 June 17, 2010 July 3, 2010 16 days 16 days
Radiometric Resolution Sensitivity of a remote sensing detector to differences in signal strength as it records the radiant flux reflected, emitted, or back-scattered from the terrain Also referred to as level of quantization BV (8-bit) 7-bit (0 – 127) 255 127 8-bit (0 –255) Grayscale 9-bit (0 –511) 12-bit (0 –1,023)
What’s Recorded Electromagnetic radiance, L, recorded in IFOV: λ = wavelength (spectral response measured in various bands or at specific wavelengths) sx,y,z = x, y, z location of the picture element and its size (x, y) t = temporal information (when, how often, and how long data were acquired) θ = set of angles describing geometric relationship among radiation source, remote sensing system, and terrain target of interest P = polarization or back-scattered energy recorded by sensor Ω = radiometric resolution (precision) at which the data (e.g., reflected, emitted, or back-scattered) are recorded by the remote sensing system Measured in watts m-2 sr-1
Not all Sensors are Alike NOAA AVHRR: 1.1 × 1.1 km (IFOV); 2,400 × 4,600 km (Swath); 14 times per day (since 1978); 4- 5 bands North and South Korea 37
Not all Sensors are Alike MODIS: 250 × 250 m to 1 km × 1km (IFOV); 10 × 2,330 km (Swath); 1-2 days – 8- and 16-day composites (since 2000); 36 bands Fires in Sinaloa, Mexico, 2001 38
Not all Sensors are Alike Landsat TM / ETM+: 15 × 15 m to 30 × 30 m (IFOV); 185 × 185 km (Swath); once every 16 days (since 1982); 7 bands Metro D.C. 39
Not all Sensors are Alike Aerial Photography: since 1930s Illinois 40
Not all Sensors are Alike Quickbird: 0.6 m × 0.6 to 2.4 × 2.4 m (IFOV); 16.5 × 16.5 km (Swath); once every 1 – 3.5 days (since 2001); 4-5 bands Giza Pyramids, Egypt 41
Not all Sensors are Alike AVIRIS: pixel size and swath depend on altitude; 20 km above: 20 × 20 m IFOV and 11 × 11 km Swath; 4 km above: 4 × 4 m IFOV and 2 × 2 km Swath; 224 bands 42
Rogan, J. and D.M. Chen. 2004. Remote sensing technology for mapping and monitoring land-cover and land-use change. Progress in Planning 61: 301-325
Rogan, J. and D.M. Chen. 2004. Remote sensing technology for mapping and monitoring land-cover and land-use change. Progress in Planning 61: 301-325
Resolutions & Applications
Earth Observation Economics Cost: consumer must be able to afford it; producer must be able to sustain / support system Goal: minimize knowledge gap between information delivery system, remote sensing experts, and information consumer (user) in a manner that is balanced with respect to cost and value (good information, easy to use, easy to understand)