Transcript Document
Electro-optical systems Sensor Resolution Outline • • • • • • • Electro-optical vs. photographic systems Spatial resolution Radiometric resolution Signal-to-noise ratio Noise equivalent reflectance, radiance, temperature Spectral Resolution Temporal Resolution Principles of Detection • Photographic camera/film systems • Digital electro-optical systems Photographic Systems vs. Electro-optical Systems • Both can be used to create images – photography is non-scanning – electro-optical systems can be either scanning or non-scanning • Both record interactions with radiation – films coated with photo-sensitive silver halide emulsions to record an image – electro-optical systems detect radiation through an electrical system • EMR detection capabilities are quite different – film is always passive, uses reflected sunlight, only works in the range of 0.4 - 1.0 mm – digital systems can be active or passive, can work in all parts of EM spectrum, high spectral resolution is possible, can be nonimaging system (e.g. laser altimeter) Digital Number at-sensor radiance imaging optics detectors electronics DN The DN that is recorded is proportional to the radiance at the sensor Digital Raster Imager Format Spatial Resolution • “A measure of the smallest angular or linear separation between two objects that can be resolved by the sensor”. (Jensen, 2000) • Resolving power is the ability to perceive two adjacent objects as being distinct – – – – – – size distance shape color contrast with background sensor characteristics • Instantaneous field of view (IFOV) is the angular field of view of the sensor, independent of height • IFOV is a relative measure because it is an angle, not a length • It can be measured in radians or degrees sensor b IFOV GIFOV • Ground-projected instantaneous field of view (GIFOV) depends on satellite height (H) and the IFOV IFOV GIFOV 2H tan 2 1 kilometer IKONOS image of Gunnison River Basin, CO 1 meter resolution 250 meter resolution SPATIAL RESOLUTION Radiometric Resolution • Number of digital values (“gray levels”) that a sensor can use to express variability of signal (“brightness”) within the data • Determines the information content of the image • The more digital values, the more detail can be expressed Radiometric Resolution • Determined by the number of bits of within which the digital information is encoded 21 = 2 levels (0,1) 22 = 4 levels (0,1,2,3) 28 = 256 levels (0-255) 212 = 4096 levels (0-4095) 8 bit radiometric resolution 2 bit radiometric resolution Dynamic Range Saturation Ideal Response (offset for clarity) Dark Current Signal Actual Sensor Response dark Scene Brightness bright Signal Strength • Need enough photons incident on the detector to record a strong signal • Signal strength depends on – Energy flux from the surface – Altitude of the sensor – Location of the spectral bands (e.g. visible, NIR, thermal, etc.) – Spectral bandwidth of the detector – IFOV – Dwell time (more on this next week) Signal-to-Noise Ratio (SNR) A sensor responds to both signal strength and electronic errors from various sensor components (noise) SNR = signal-to-noise ratio signal noise signal = the actual energy reaching the detector noise = random error in the measurement (all systematic noise has been removed) must have high SNR To be effective, sensor Noise DN n i1 m DN i 2 n 1 mDN is the mean value of the DNs in the sample population n is the number of DN values in the sample population 50% A uniform material of known reflectance Mean DN = 201 Noise = 1.345 SNR = 201/1.345 = 149 200 201 199 203 202 201 200 DNs from a single detector, measured in a laboratory (prelaunch) Noise Equivalent Radiance Noise Equivalent Reflectance Noise Equivalent Temperature • A measure of the smallest magnitude of signal or of a change in signal that can be detected • Can be expressed in terms of radiance (L) or reflectance (r) or temperature (T) Converting SNR to NEr: Divide the noise value by the DN value recorded over a 100% reflectance target Since the sensor reads a value of 201 over a 50% reflectance target we assume that it would read 402 over a 100% reflectance target NEr = 1.345 402 = 0.0033 = 0.33% 50% Noise DN n i1 m DN i 2 n 1 mDN is the mean value of the DNs in the sample population n is the number of DN values in the sample population Noise Equivalent Radiance Noise Equivalent Reflectance Noise Equivalent Temperature • “Noise equivalent” is a measure of the smallest magnitude of a real change in signal that can be detected • Can be expressed in terms of radiance (L) or reflectance (r) or temperature (T) Spectral Resolution • The width and number of spectral intervals in the electromagnetic spectrum to which a remote sensing instrument is sensitive • Allows characterization based on geophysical parameters (chemistry, mineralogy,etc.) Spectral Resolution • Determined by: – the number of spectral bands – width of each band • Described by the full-width at half-maximum (FWHM) • spectral response function (SRF) of each band Multiple bands: • Surface components with very distinct spectral differences can be resolved using broad wavelength ranges Hyperspectral Remote Sensing Hundreds of bands Temporal Resolution • The frequency of data acquisition over an area • Temporal resolution depends on: – the orbital parameters of the satellite – latitude of the target – swath width of the sensor – pointing ability of the sensor • Multi-temporal imagery is important for – infrequent observational opportunities (e.g., when clouds often obscure the surface) – short-lived phenomenon (floods, oil spills, etc.) – rapid-response (fires, hurricanes) – detecting changing properties of a feature to distinguish it from otherwise similar features TEMPORAL RESOLUTION Examples of sensor temporal resolution: SPOT - 26 days (1, 4-5 days with pointing) Landsat - 16 days MODIS - 16 day repeat, 1-2 day coverage AVHRR – 9 day repeat, daily coverage GOES - 30 minutes More about this when we discuss orbits