1 RADIOMETRIC CALIBRATION Lecture 10
Digital Image Data 2 Digital data are matrices of digital numbers (DNs) There is one layer (or matrix) for each satellite band Each DN corresponds to one pixel
3
Digital Image Data 4 Images are presented as 2-d arrays. Each pixel (array element) has a location (x,y) and associated with it a digital number (DN). Position of pixel often describe in terms of rows and columns 0 0 0 2 0 2 3 4 F(2,3) F(4,1) 2 2 6 9 2 2 9 12
What are digital numbers (DNs)? 5 DNs are relative measures of radiance DNs are NOT reflectance DNs can be converted to ground reflectance if we know atmospheric properties, etc. The range of DNs depends on the radiometric resolution of the instrument
Landsat Group 6 Rescaling factors for converting calibrated Digital Numbers (DNs) to absolute units at-sensor radiance The Landsat series of satellite provides longest continuous record of satellitebased observation Landsat satellite can be classified into 3 groups based on sensor and platform characteristics : Gr. 1 Landsat 1, Landsat 2, Landsat 3 with the Multispectral Sacnner (MSS) sensor and the Return Beam Vidicon (RBV) camera Gr. 2 Landsat 4, Landsat 5 which carry the Thematic Mapper (TM) as well as the MSS sensor Gr. 3 Landsat 6 and Landsat 7 which include the Enhanced Thematic Mapper (ETM) and the Enhanced Thematic Mapper Plus (ETM+) sensors respectively
Radiometric calibration 7 The ability to detect and quantify changes on the Earth s environment depends on the sensor that can provide Calibrated and consistent measurement of the Earth s surface The correct interpretation of scientific information from a global, long-term series of RS products requires the ability to discriminate between product artifacts and changes in the Earth processes being monitored.
Radiometric calibration 8 In the application of satellite RS data for quantitative investigations, the first and foremost important step are : Calibration of satellite data Atmospheric correction Topographic correction Spectral signatures plays an important role in detecting and analyzing Earth surface Different material types can be spectrally similar making identification and classification difficult Through understanding the limitations and constraints of RS system and its related products is of vital importance
Radiometric calibration 9 Radiometric correction applied to digital data varies significantly among sensors. Radiance measured of any object by a RS is influenced by : Bi-directional reflectance distribution (BRDF) Changes in scene illumination Atmospheric attenuation Viewing and topographic geometry Instrument response characteristics
Radiometric calibration 10 Various paths of irradiance and surface reflectance can alter the apparent at-sensor radiance of a given target Path radiance Total at-sensor radiance : Diffused radiation Reflected radiation Emitted radiation
Conversion to at-sensor spectral radiance (Q cal -to-l λ ) 11 Calculation of at-sensor spectral radiance is the fundamental step in converting image data from multiple sensors and platforms into a physically meaningful common radiometric scale Radiometric calibration of the MSS, TM and ETM+, sensors involves rescaling the raw digital numbers (Q) transmitted from the satellite to calibrated digital numbers (Q cal )
Conversion to at-sensor spectral radiance (Q cal -to-l λ ) 12 During calibration the pixel values (Q) from raw, unprocessed image data are converted to units of absolute specatral radiance using 32-bit floating point calculation The absolute values are then scaled to Q cal before media output: 7-bits (MSS, Q calmax =127) 8-bit (TM and ETM+, Q calmax =255) 16-bit (ALI, Q calmax =32767)
Conversion to at-sensor spectral radiance (Q cal -to-l λ ) 13 Conversion from Q cal back to at-sensor spectral radiance (L λ ) requires knowledge of the lower and upper limit of the original rescaling factors To convert from Q cal back to L λ then requires knowledge of the original rescaling factors (LMIN λ and LMAX λ ) which have changed over time. Different scaling factors for Landsat 5 and 7
Conversion to at-sensor spectral radiance (Q cal -to-l λ ) 14
Conversion to at-sensor spectral radiance (Q cal -to-l λ ) 15 Landsat 7, Landsat ETM+ gain data that the satellite can operate under two different gain states (unlike Landsat 5). Gain selection for a scene is controlled by the Mission Operation Center The goal is for maximizing the instrument's 8 bit radiometric resolution without saturating the detectors
16 ETM+ Spectral Radiance Range in watts m -2 ster -1 μm -1 Processed Before July 1, 2000 Proccessed After July 1, 2000 Low Gain High Gain Low Gain High Gain LMIN LMAX LMIN LMAX LMIN LMAX LMIN LMAX Band Number 1-6.2 297.5-6.2 194.3-6.2 293.7-6.2 191.6 2-6.0 303.4-6.0 202.4-6.4 300.9-6.4 196.5 3-4.5 235.5-4.5 158.6-5.0 234.4-5.0 152.9 4-4.5 235.0-4.5 157.5-5.1 241.1-5.1 157.4 5-1.0 47.70-1.0 31.76-1.0 47.57-1.0 31.06 6 0.0 17.04 3.2 12.65 0.0 17.04 3.2 12.65 7-0.35 16.60-0.35 10.932-0.35 16.54-0.35 10.80 8-5.0 244.00-5.0 158.40-4.7 243.1-4.7 158.3
Conversion to at-sensor spectral radiance (Q cal -to-l λ ) 17 The gain state for a scene is matched to the expected brightness conditions. For all bands, the low gain dynamic range is approximately 1.5 times the high gain dynamic range hence imaging in low gain mode when surface brightness is high and in high gain mode when surface brightness is low Prior to July 13th, 2000, band 4 always operated in high gain mode when imaging land (land is classified as non-desert and non-ice). After July 13th, 2000, low gain mode was used when sun elevation exceeded 45 and high gain mode continued to be used for land when the sun elevation did not exceed 45
18
ETM+ data characteristics: gain setting 19 Each Landsat-7 Path/Row location is categorized into one of the following six types - for each surface cover type the gain setting rules are different: 1. Land (non-desert, non-ice) 2. Desert 3. Ice/Snow 4. Water 5. Sea Ice 6. Volcano/Night Land (non-desert, non-ice): Bands 1-3 set to high gain Band 4 set to high gain except where sun elevation is greater than 45 to avoid dense vegetation (reflectance > 0.66 ) saturation. (At this sun angle high gain in band 4 saturates at about a reflectance of 0.66, so switching to low gain keeps targets at this reflectance or below from saturating.) Bands 5,7 set to high gain Band 8 set to low gain
ETM+ data characteristics: gain setting 20 Desert: Bands 1-3 set to high gain except where sun elevation is greater than 28 - to avoid bright desert target (reflectance >0.65 in band 3, >0.66 in band 1, >0.71 in band 2) saturation Band 4 set to high gain except where sun elevation is greater than 45 (set to low gain) - to avoid bright desert (reflectance > 0.66 ) saturation. Band 5,7 set to high gain except where sun elevation is greater than 38 -- to avoid bright desert target (reflectance >0.70 in band 5, [>0.68 in band 7] saturation Band 8 set to low gain
ETM+ data characteristics: gain setting 21 Ice/Snow and Sea Ice : Bands 1-3 set to high gain except where sun elevation is greater than 19 - to avoid snow ice (reflectance > 0.95 in band 3, >0.94 in band 1, >1.03 in band 2) saturation. Band 4 set to high gain except where sun elevation is greater than 31 - to avoid snow/ice (reflectance >0.92) saturation. Band 5, 7 set to high gain Band 8 set to low gain Water/Coral Reefs : Bands 1-5,7 set to high gain Bands 8 set to low gain Volcano/Night night time imaging (sun elevation < 0) is only routinely performed for sights identified as "Volcano : Bands 1-4, set to high gain Bands 5,7 set to low gain to reduce saturation of volcanic hot spots Band 8 set to low gain
ETM+ data characteristics: gain setting 22
Conversion of radiance to reflectance (L λ to- ρλ) 23 For relatively clear Landsat scenes, a reduction in between-scene variability can be achieved through a normalization for solar irradiance by converting spectral radiance, as calculated above, to planetary reflectance or albedo. This combined surface and atmospheric reflectance of the Earth is computed with the following formula: 2. L. d p Where: E SUN.cos s p E SUN d L θ s = Unitless planetary reflectance = Mean solar exoatmospheric irradiances (Global flux) [W/(m 2 μm)] = Earth-Sun distance in astronomical units = Spectral radiance at the sensor's aperture [W/(m 2 sr μm)] = Solar zenith angle in degrees
Day of Year Distance Day of Year Distance Day of Year Distance Day of Year Distance Day of Year Distance 24 1.98331 74.99446 152 1.01403 227 1.01281 305.99253 15.98365 91.99926 166 1.01577 242 1.00969 319.98916 32.98536 106 1.00353 182 1.01667 258 1.00566 335.98608 46.98774 121 1.00756 196 1.01646 274 1.00119 349.98426 60.99084 135 1.01087 213 1.01497 288.99718 365.98333
Conversion to at-sensor brightness temperature (L λ -to- T) 25 The thermal band data (Band 6 on TM and ETM+) can be converted from at-sensor spectral radiance to effective at-sensor brightness temperature. The at-sensor brightness temperature assumes that the Earth s surface is a black body (i.e., spectral emissivity is 1), and includes atmospheric effects (absorption and emissions along path). The at-sensor temperature uses the prelaunch calibration constants
Conversion to at-sensor brightness temperature (L λ -to- T) 26 The conversion formula from the at-sensor s spectral radiance to at-sensor brightness temperature is:
Conversion to at-sensor brightness temperature (L λ -to- T) 27 The ETM+ Level 1 product has two thermal bands: one acquired using a low gain setting (often referred to as Band 6L; useful temperature range of 130 350 K) other using a high gain setting (often referred to as Band 6H; useful temperature range of 240 320 K).