Remote Sensing Exam 2 Study Guide

Transcription

1 Remote Sensing Exam 2 Study Guide Resolution Analog to digital Instantaneous field of view (IFOV) f ( cone angle of optical system ) Everything in that area contributes to spectral response mixels Sampling rate Determines horizontal resolution # of sampling levels determines bits Instantaneous field of view (IFOV) f ( cone angle of optical system ) Everything in that area contributes to spectral response mixels Dwell time the time a sensor spends looking at any area Analog-Digital conversion process Photons strike detector, electrons flow proportional to the # photons # photons related to reflectance characteristics, λ (recall that shorter wavelengths carry more energy) Area viewed (more photons come from a larger area (at any λ)) Spectral region width (wider region = more photons) Sampling rate Determines horizontal resolution # of sampling levels determines bit Spatial resolution Spectral resolution Radiometric resolution There are tradeoffs If you increase the spatial resolution (i.e. smaller area of coverage), detector is receiving energy from a smaller area (less total energy received at detector at some λ) So, the spectral range needs to be increased to receive enough energy (or design a better detector) If you increase the spectral resolution (i.e. more bands each representing a smaller range of the EMS) detectors are receiving energy in a narrower range of the spectrum (less total energy received at the detector at some λ) So, the DWELL time needs to be increased to compensate, resulting in lower spatial resolution (or design a better detector) Coverage area (altitude and field of view) Off-nadir capabilities (point sensor to side or front/aft)

2 Pushbroom System Advantages Fewer moving parts Less energy drain Less weight Longer dwell time Better geometric fidelity Disadvantages More sensors to calibrate Limited spectral range Orbital characteristics of satellites Altitude Period Inclination Equatorial crossing Sun synchronous Combination of orbital period and inclination that cause the satellite to cross the equator at the same time (keeps pace with the sun s westward progress Geostationary Equatorial orbit of an altitude that results in an orbital period of 24 hours. 36,000km Earth resources space imaging ERTS-1 Earth Resources Technology Satellite A feasibility test First launched 7/23/1972 But design began 1967 Planned to launch six satellites Pre-launch designated as ERTS A,B,C,D,E,F Post-launch redesignated as ERTS 1,2,3,4,5,6 Open-sky principle All countries may evaluate Just before ERTS-B launch, NASA renamed the program to LANDSAT To differentiate it from a future program launch Seasat ERTS1 renamed Landsat1 retroactively

3 Landsat 1,2, 3 Similar characteristics 185km swath width Sun synchronous orbit 9 o inclination to equator Successive orbits 2760km apart 104 minutes orbital period Crosses equator at 9:42AM local time Takes advantage of normally better atm. cond early in day Also provides same illumination conditions for comparing images Large gaps between orbits on a given day (14/day) 18 day temporal resolution Overlap at equator 14%, at poles 84% Return-beam Vidicon (RBV) system RBV sensed in 3 bands equivalent to CIR film G,R,IR Bands 1,2,3 Television camera-like Used a photosensitive surface with shutter, then scanned Instant image (like camera) better geometric fidelity Spatial resolution 80m Landsat 3 changed to improve spatial resolution (30m) also single band ( um) 2 cameras (side-side) MSS G, R, NIR, NIR.5.-6,.6-.7,.7-.8,.8-1.1um Bands 4,5,6,7 Landsat 3 and B Failed 6 contiguous lines scanned IFOV = 79m A-D on-board 0-63 (6-bit) B4-6 rescaled to (7-bit) On ground Sampling rate = /s Results in 56m horizontal spacing

4 Landsat 4, day orbital return period (i.e. the temporal resolution) Altitude 705km 185km swath width Sun synchronous orbit 98.2 o inclination to equator Crosses at 9:45Am Successive orbits 2760km apart 104 minutes orbital period Also provides same illumination conditions for comparing images Orbits (14.5/day) Overlap at equator 14%, at poles 84% 8 bit resolution Except band 6 is 7 bit 30m IFOV Except band 6 is 120m on TM4,5 60m on TM7 15m Pan on TM7 16 detectors for all bands (4 for B6) totaling 100 detectors Note: Landsat 7 Added a pan Band 8 15m res0.4 to 0.9um

5 SPOT satellite French satellite (Toulouse, France) SPOT 1,2,3 26 day repeat period 20 m res. MSS 3 bands (G,R, NIR) 10 m res. Pan SPOT 4 launched March 1998 HRVIR Includes Mid-IR band (20m) Pan band replaced by red band Vegetation instrument 2250 km swath, 1 km IFOV, same bands as HRVIR but blue used instead of green For oceanographic appllications SPOT 5 - launched May m res. (pan HRGeometric) May be resampled to 2.5 m 10 m res G,R,NIR 20 m res MIR (due to CCD limitations) HRStereoscopic instrument Fore-aft instruments for DEM generation Global 10 m AVHRR Advanced Very High Resolution Radiometer

6 EOS ESE Science Objectives Provide the first state distribution of the main Earthatmosphere coupled parameters Improve our ability to detect human impacts on climate, identify fingerprints of human activity on climate, and predict climate change Provide observations that will improve forecasts of the timing and geographical extent of transient climatic anomalies Improve seasonal and inter-annual predictions Develop technologies for disaster prediction, characterization, and risk reduction from wildfires, volcanoes, floods, and droughts. Start long-term monitoring of the change in global climate and environmental change. EOS (ESE) AM-1 Mission Overview Terra and Aqua Launch date: December 1999 Terra May 2002 Aqua Orbit: 705 km altitude, polar Orbit period: 98.8 minutes Equator crossing: 10:30 AM descending Terra 1:30 PM Ascending Aqua Ground track repeat cycle: 16 days Terra Instruments: Moderate Resolution Imaging Spectroradiometer (MODIS) Advanced Spaceborne Thermal Emission Radiometer (ASTER) Multi-angle Imaging Spectroradiometer (MISR) Measurement of Pollution in the Troposphere (MOPITT) Clouds and the Earth Radiant Energy System (CERES)

7 MODIS 12-bit radiometric Resolution for all Bands 2-day global coverage Excellent band-band Registration and Radiomertric accuracy Aster 3 unique instruments (has off-nadir capabilities) Visible and Near Infrared (VNIR) 3 bands on nadir (bands 1-3) G, R, NIR 1 band 27.5º rear-looking (NIR & same as B3 on nadir) Capable of DEM generation 15m res Short Wave Infrared (SWIR) 6 bands (bands 4-9) 30m res Thermal Infrared (TIR) 5 bands (bands 10-14) 90m res DIGITAL IMAGE PROCESSING Image rectification & restoration Correct distortions and degradations to imagery Geometric distortions Radiometric distortions Sensor dependent Together, called preprocessing techniques Image Enhancement More effectively portray image data for visual interpretation Many techniques No set best way Trial and error Sometimes several enhancements to a single image are the best way

8 Image classification Quantitative techniques for automating the identification of features in a scene Multispectral data Statistical based decision rules Spectral pattern recognition LULC mapping Data merging / GIS Change detection Merging with GIS (LULC) w/zoning, topo, etc Multisensor merging Multitemporal data merging Hyperspectral image analysis Dozens to hundreds of bands Biophysical modeling Crop yield, water depth, insect infestation, pollution, etc Image compression Statistics Mean Standard Deviation Histograms Graphical distribution of values in single band Used for interpretation And for enhancements Scatterplot & Ellipses Used for interpretation of band-pairs Enhancements Process of making an image more interpretable Technique used f(dataset, desired result) Must know characteristics of dataset Have an objective Ex. Sharpening a dataset to better delineate boundaries Ex. Reducing the number of bands May be permanent or on-screen only Spectral enhancements Deals with pixels in different bands Spatial enhancements Deals with surrounding pixels in a single band Derive a new value based on values in surrounding pixels Moving window concept

9 Image enhancements Contrast enhancements Gray-level thresholding Level slicing Contrast stretching Spatial feature manipulation Filtering Edge enhancement Fourier analysis Multi-image manipulation Band ratioing Principle components IHS color space transforms Vegetation components Contrast enhancements A form of spectral enhancements Increases the contrast in certain spectral ranges of the image Likely at the expense of others Goal to make image more interpretable or features more identifiable Application in one band may not be appropriate for others (each band handled separately) Contrast Stretching Linear contrast stretch Simple Sinusoidal stretch Divides histogram into several user-defined parts Doesn t eliminate detail in some parts of image Histogram Equaliztion LUT values assigned based on frequency of occurrence Large regions of LUT reserved for common DN Small regions of LUT reserved for infrequent DN Concept based upon information yield Greatest information in most frequent pixels Special stretches Enhance whatever you re interested in Water, veg, etc Contrast enhancements Grey-level thresholding Segment image into 2 classes One above/below some user-defined value Often used to prepare a binary mask Level slicing Divide histogram into segments Each segment receives the same DN Each coded (colors) Elevations, Thermal imagery

10 Destriping Sixth line striping or Variations in calibration (sensitivity) of same sensor on different lines Produces contrast variation parallel to scan Fix Averaging Histogram Normalization MUST be done prior to geometric correction