4 Learning ObjectivesWhat are the four types of resolution that we must consider with remotely sensed data?Be able to define each type of resolution.Be able to calculate the number of pixels in a given area.Understand the trade-offs between different types of resolution.Understand the relationship between SNR and resolution.
5 Learning Objectives (cont.) Understand binary data and the relationship between radiometric resolution and storage space.Understand the difference between different types of orbits.
6 What are the four types of resolution? SpatialSpectralRadiometricTemporal
7 Spatial ResolutionUsually reported as the length of one side of a single pixelIn analog imagery, the dimension (e.g. width) of the smallest object on the ground that can be distinguished in the imageryDetermined by sensor characteristics (for digital imagery), film characteristics (for air photos), field of view, and altitude.
8 Group ProblemIf you have a study area that covers 1 km2, how many Landsat 30 m pixels does it take to cover it (nearest whole number)?How many 15 m panchromatic pixels would it take to cover the same area?
16 Spatial Resolution Trade-offs Data volumeSignal to Noise Ratio“Salt and Pepper”Cost
17 Spectral ResolutionCan be described two ways, but they usually go hand in hand.How many spectral “bands” an instrument recordsHow “wide” each band is (the range of wavelengths covered by a single band)
18 Spectral resolution Related to the measured range of EMR Wide range - coarser resolutionNarrow range - finer resolution
19 Case 1 Measure the EMR across a wide range E.g., a single panchromatic band covering the entire visible portion of the spectrumAssigns a single DN representing all visible light energy hitting the sensorAnalogous to black and white (panchromatic) film
26 High Spectral Resolution Spectral reflectance curve for green leaf using 224 bands (high spectral resolution)ReflectanceWavelength (nm)Low Spectral ResolutionSpectral reflectance curve for green leaf using 6 bands (lower spectral resolution)ReflectanceWavelength (nm)
27 Could you distinguish Dolomite from Calcite using Landsat 8 spectral data?
28 Spectral Resolution Trade-Offs Data Volume and processing1 DN for each pixel in EACH BANDSignal to Noise RatioCost
29 Group ProblemFor your 1 km2 study area, if you use 7 Landsat 8 bands, how many DNs will your computer have to store?
30 Radiometric Resolution How finely does the satellite divide up theradiance it receives in each band?How much light does it take to changethe DN from one number to the next?Usually expressed as number of bits used to store the maximum possible DN value8 bits = 28 = 256 levels (usually 0 to 255)16 bits = 216 = 65,536 levels (0 to 65,535)
32 Radiometric resolution 1 bit ( 0 - 1)8 bit ( ) (older Landsats, many others)16 bit ( ,535 ) (Landsat 8)32 bit ( 0 - 4,294,967,295 ) (uncommon)For an 8-bit satellite:DN = 0: No EMR or below some minimumamount of light (threshold)DN = 255: Max EMR or above some maximumamount of light
33 Radiometric resolution 8 bit data (e.g., Landsat 5) (256 values)Everything will be scaled from 0 – 255Subtle details may not be represented16 bit data (e.g., Landsat 8) (65,536 values)Wide range of choicesRequired storage space will be twice that of 8 bit
34 Radiometric Radiation Trade Offs Data volumeEvery 8 bits takes 1 byte to store ona computer.One 8-bit DN takes 1 byteOne 16-bit DN takes 2 bytesEtc.
35 Group ProblemIf your are using 7-band, 16-bit Landsat 8 data for your 1 km2 area, how many bytes are needed to store your DNs on your computer?
36 Calculating Image Size Computer hard drives store data in “boxes” called bytes (e.g., 1 Mb = 1 million bytes) 1 byte can hold 8 binary (base 2) digits (0s or 1s or some combination of 0s and 1s) Each “bit” is a single binary digit An 8-bit number is made of of 8 binary digits and fits into 1 byte. A 9-bit number won’t fit in 1 byte and requires 2 bytes.
37 Converting Base 10 to Binary Base 2 (Binary)1210311410051016110711181000255256257(etc.)
38 Temporal resolution Time between two subsequent data acquisitions for an areaAll of the Landsat satellites have a 16-day return timeMODIS has a 1-2 day return time.
39 Return Time (Temporal Resolution) Depends on:Orbital characteristicsSwath widthAbility to point the sensor
40 Orbital Characteristics GeosynchronousPolarSun synchronous
41 Geosynchronous Orbits Satellite orbits the earth at a rate that allows it to match the earth’s rotation—so the satellite is always over the same placeNarrow range of altitudes—about 35,786 km above the equator.Useful for communications, weather etc.Example: GOES satellite (weather)Geosynchronous orbiting earth satellite
42 Polar/Sun Synchronous Orbits Pass roughly over the north and southpolesFly over the same place on earth at thesame time of dayExamples: Landsat, AVHRRGood for land remote sensingReturn time related to spatial resolution,latitude, swath width, and orbital altitude
43 Return Time Trade Offs Spatial resolution Viewing geometry effects (off nadir)Clouds and other atmospheric problemsLack of archival repeat coverage forpointable satellites
44 In summary, choosing a satellite is often an exercise in weighing the relative trade-offs of resolution against data needs (and budgets!).