1. 1“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Dr. Franz J Meyer, Dr. Rüdiger Gens Earth & Planetary Remote Sensing & Alaska Satellite Facility, University of Alaska Fairbanks Geocoding & Mosaicking

2. 2“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Outline Coordinate Systems and Projections Geocoding Terrain Correction –Geometric Terrain Correction –Radiometric Terrain Correction Data Fusion & Mosaicking N

3. 3“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Coordinate Systems and Projections N

4. 4“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Geographic latitude Geodetic latitude a b f Coordinate systems Geographic coordinates a: semi-major axis b: semi-minor axis f: flattening = (a-b)/a Expresses as a fraction 1/f = about 300 –geographical coordinates imply spherical Earth model –geodetic coordinates imply ellipsoidal Earth model

5. 5“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED How to describe the Globe? assumption that the earth is a sphere is possible for small-scale maps (smaller than 1:5000000) to maintain accuracy for larger-scale maps (scales of 1: 1000000 or larger) a spheroid is necessary Source: ArcGIS help file

6. 6“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Common Spheroids Bessel 1841 Clarke 1866, Clarke 1880 GEM 6, GEM 10C GRS 1967, GRS 1980 International 1924, International 1967 WGS 72, WGS 84

7. 7“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED three reference surfaces –ellipsoid –geoid –topography Reference surfaces Source: R. Gens, UAF

8. 8“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED ellipsoid defines mathematical surface approximating the physical reality while simplifying the geometry Reference surfaces Source: R. Gens, UAF

9. 9“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED geoid defined as level surface of gravity field with best fit to mean sea level –maximum difference between geoid and mean sea level about 1 m –Up and down means water is flowing! Reference surfaces Source: R. Gens, UAF

10. 10“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED topography represents the physical surface of the Earth Reference surfaces Source: R. Gens, UAF

11. 11“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED describes the relationship between a particular local ellipsoid and a global geodetic reference system (e.g. WGS84) local datum defines the best fit to the Earth's surface for particular area (e.g. NAD27) Datum Source: ESRI

12. 12“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Common Datums World Geodetic System 1972 (WGS 72) World Geodetic System 1984 (WGS 84) North American Datum 1927 (NAD 27) North American Datum 1983 (NAD 83) European Datum 1950 (ED 50) South American Datum 1969 (SAD 69)

13. 13“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Geographic coordinate system A point is referenced by its longitude and latitude values Longitude and latitude are angles measured from the earth’s center to a point on the earth’s surface Source: ESRI

14. 14“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Cartesian Coordinates geodetic coordinates inappropriate for satellite imagery  cartesian coordinates z y x P

15. 15“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED problem of mapping three-dimensional coordinates related to a particular datum on a flat surface –maps are two-dimensional –impossible to convert spheroid into flat plane without distortions of shape, area, distance, or direction  map projections Map projections

16. 16“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Cylindrical projections cylinder that has its entire circumference tangent to the Earth’s surface along a great circle (e.g. equator)

17. 17“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED The map projection has distorted the graticule (data near the poles is stretched) Cylindrical projections Source: ESRI

18. 18“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Cylindrical: Examples Mercator projection Transverse Mercator projection Oblique Mercator projection Source: ESRI

19. 19“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Conic projections Simplest conic projection is tangent to the surface along a small circle (called standard parallel). The meridians are projected onto the conical surface, meeting at the apex. Parallel lines of latitude are projected onto the cone as rings.

20. 20“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Conic projections further you get from the standard parallel, the more distortion increases. Source: ESRI

21. 21“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Conic projections secant projection: a more complex conic projection contact the global surface at two locations defined by two standard parallels less overall distortion than a tangent projection Source: ESRI

22. 22“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Conic: Examples Conic projection with two standard parallels Lambert Conformal Conic projection (preserves angles) Albers Conic Equal-Area projection (preserves areas)

23. 23“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Azimuthal (Planar) projections –projecting positions directly to a plane tangent to the Earth’s surface

24. 24“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Azimuthal (Planar) projections Source: ESRI Point of contact specifies the aspect and is the focus of the projection. The focus is identified by a central longitude and a central latitude. Possible aspects are polar, equatorial, and oblique. Examples: –Lambert Azimuthal Equal-Area projection –Stereographic (conformal) projection

25. 25“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Image Examples Cylindrical projection Conical projection http://www.satellite-images.com/ http://www.fes.uwaterloo.ca/crs/geog165/conproj.htm

26. 26“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Geocoding N

27. 27“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Definitions Geocoding –geometric transformation of an image into a cartographic map projection Georeferencing –relating image coordinates to map coordinates by defining control points (usually image corners) Geometric correction and image rectification are sometimes used synonymously –geocoding maybe part of geometric correction

28. 28“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Geocoding by co-registration Image to image –reference needs to be map projected Image to map –map in raster or vector format –map needs to have map coordinates Image with measured ground control points –ground control points (GCPs) need to be identified in the image –GCPs need to be known in some map coordinate system

29. 29“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Sensor Geometric Model Sensor Model –sensor specific –analytical reconstruction of image formation using orbit and sensor parameters –corrects image globally –small number of ground control points to improve parameters –DEM Z X Y x Greenwich S i o w v N r

30. 30“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Geocoding steps Relation between image coordinates and geographic coordinates using image and sensor geometry and DEM information –line / sample  latitude / longitude Conversion of geographic coordinates into map projected coordinates –latitude / longitude  x map / y map –choice of map projection and datum Determination of a transformation function to map image coordinates into projection coordinates –usually quadratic, at times cubic –linear least squares polynomial fit Resampling using mapping function –determination of pixel value in the map projected using one of the interpolation methods

31. 31“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Standard interpolation methods Nearest neighbor interpolation –takes pixel value closest to calculated location –preserves original pixel values Bilinear interpolation –weighted average (2x2 kernel) –smoothing effect Cubic convolution –third degree polynomial fit (4x4 kernel) –essentially low-pass filter x x x

32. 32“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Example: Original image

33. 33“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Example: Transformed image

34. 34“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Example: Geocoded image

35. 35“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Terrain Correction N

36. 36“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Terrain Correction Terrain Correction “orthorectifies” SAR data. Data is resampled so that pixels appear in the proper geolocation. One can overlay SAR data onto remote-sensing data from different sensors and/or geometries.

37. 37“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Geometric Terrain Correction remove effects of side looking geometry of SAR images necessary step to allow geometric overlays of remotely sensed data from different sensors and/or geometries

38. 38“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Geometric Terrain Correction Methods Backward geocoding –determine time shift to match DEM and SAR image –correction in map projected space –"paint the DEM" Forward geocoding –determine offset to match DEM and SAR image in SAR image geometry –correction in image space (slant range geometry) –geocoding can be separate step

39. 39“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Backward geocoding "painting" the DEM in map projection space Starting from a predefined pixel location in the geocoded output image, the corresponding column and line in the slant-range image are calculated using the DEM information. The gray value for the output pixel is calculated through interpolation in the slant range image.

40. 40“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Backward geocoding Backward geocoding scheme: –Ground pixels of different geographic location may be projected onto same slant-range pixel due to layover and foreshortening. –Pixels of varying slant range coordinate may have the same geographic location (e.g. think of a vertical building). Small, Remote Sensing Lab, University Zurich

41. Forward geocoding Forward geocoding starts from a pixel in the slant range image for which the location on ground is determined. This results in an irregularly distributed point cloud on ground, and a subsequent resampling step is necessary. Terrain correction is done in image space. This method is implemented in the MapReady Software. 41

42. 42“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Layover / Shadow masks can be derived from DEM useful to provide information about problem areas –shadow regions – no information available –layover and foreshortening – reduced spatial resolution

43. 43“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Layover / shadow masks overlay of the layover mask on the terrain corrected image –green: layover –red: shadow –blue: user mask –dark grey: invalid data

44. 44“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Radiometric Terrain Correction some SAR applications require absolute radiometric calibration accurate to within 1 dB –e.g. biomass estimation →requires generalization of many assumptions widely made in the SAR literature –radar equation –area effect

45. 45“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Radiometric terrain correction radiometric correction (applied on the right) ensures that the original contrast of the image is preserved

46. 46“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Terrain Correction SAR Image With layover Terrain Corrected Image Without layover

47. 47“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Geocoding Terrain corrected image before geocoding After geocoding

48. 48“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Data Fusion and Mosaicking N

49. 49“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Data Fusion Landsat R1 - Terrain Corrected Fused Product Terrain Correction permits SAR co-registration with other datasets.

50. 50“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Terrain Correction Image terrain corrected using ASF MapReady Tool

51. 51“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Data Fusion PALSAR image fused with AVNIR-2

52. 52“Principles & Applications of SAR” Instructor: Franz Meyer © 2009, University of Alaska ALL RIGHTS RESERVED Image Mosaicking Goal: Compose seamless large-area map from images of small spatial extent Two step procedure: –Image registration –Image composition Image registration: –Relative alignment of images and removal of geometric distortions –Projection into common reference system Image composition: –Definition of seam lines –Histogram matching in overlap areas –Blending functions

53. Image Registration Example datasets: Brasil (Rio de Janeiro) –Image projection into same system → registration 53

54. Definition of Cut Lines Example datasets: Brasil (Rio de Janeiro) –Cut lines are lines in the overlap area of images along which a fusion of the images will be performed –Definition of cut lines either manually or automatically –Automatic definition based on geometry and pixel values 54

55. Definition of Cut Lines Example datasets: Brasil (Rio de Janeiro) –Cut lines are lines in the overlap area of images along which a fusion of the images will be performed –Definition of cut lines either manually or automatically –Automatic definition based on geometry and pixel values 55

56. Blending Function & Mosaicking Example datasets: Brasil (Rio de Janeiro) –Blending function may include Histogram matching in overlap areas Smoothing close to cut lines Feathering to hide slight registration errors –Mosaicking creates final image 56

57. Final Mosaic Example datasets: Brasil (Rio de Janeiro) 57