Mars Data Analysis Using IDL GG7101 Lecture 11

1
Mars Data Analysis Using IDLGG710-1Lecture 11

• F. Scott Anderson
• Office POST 526B
• Phone [phone]
• Email anderson_at_higp.hawaii.edu

2
Review of 11/8/04

• Bar plots, plotting error, histograms
• Discuss thermal emission, 2D Histograms,
  homework
• Irregular gridding contour, mesh plots

3
Today

• Quiz 8 returned / Homework NOT returned
• Questions from last time
• Quiz 9
• Finish Graphics
• Shaded surface plots, large images
• Copying a screen image
• Reading and writing graphics files
• Other devices postscript and z-buffer
• Curve Fitting
• Polynomial least square fits
• Minimum absolute deviation fits
• Review of today's class

4
Questions from 11/8/04

• Stay, hit, or fold?

5
Quiz 8 Scores
6
Quiz
7
Quiz Answers
8
Quiz Answers
9
Shaded Surface Plots

• It is possible to produce a 3D perspective solid
  surface image using SHADE_SURF
• Form
• SHADE_SURF, z_img,x,y, AZvar1, AXvar2,
  SHADEimg
• Can use any image that is scaled from 0 to 255
  and that has the same size as z_img
• Thus can make 3D topo map with image draped over
  it

10
Example Shaded Surface Plots

• restore, 'test.data'
• SHADE_SURF, z, x, y, shadebytscl(dist(50))

11
Example Shaded Surface Plots

• Bad Viking image draped on MOLA topo of hematite
  region

12
Large images SLIDE_IMAGE

• Large images are difficult to show on the screen
• If you want to maintain full resolution, you may
  desire a scrolling window and a shrunken image
• Form
• SLIDE_IMAGE, Image, FULL_WINDOWvariable,
• SLIDE_WINDOWvariable, SHOW_FULL0,
  TITLEstring,
• XVISIBLEwidth, YVISIBLEheight, XSIZEwidth,
• YSIZEheight
• Where
• Image is optional
• FULL_WINDOW and SLIDE_WINDOW are used with WSET
• SHOW_FULL0 results in only a scrolling window
• XVISIBLE/YVISIBLE are the size of the windows
• XSIZE/YSIZE is the size of the scrolling image

13
SLIDE_IMAGE Examples

• xdist(1000,1000)
• slide_image, x

14
SLIDE_IMAGE Examples

• xdist(1000,1000)
• slide_image, x, show_full0

15
SLIDE_IMAGE Examples

• xdist(1000,1000)
• slide_image, full_windowfw, slide_windowsw,
• xvisible600, yvisible600, xsize1000,
  ysize1000
• wset, sw
• tvscl, x

16
Reading Images from a Device

• If you have built a complex image composed of
  many plot, tvscl, and contour calls, you may want
  to read that image and store it
• Fundamentally limited by the number of pixels on
  the screen
• Useful for writing bitmap or raster images, or
  images composed of pixels (instead of vector
  images which use lines, circles, etc)
• Can be done using TVRD function
• Form
• Result TVRD(/ORDER)
• On a 24 bit display
• Result TVRD(/TRUE,/ORDER)
• IDL can write read many types of bitmap images
  (tif, jpg, pict, bmp, pnm, etc)

17
Example TVRD an Image

• restore, 'test.data'
• SHADE_SURF, z, x, y, shadebytscl(dist(50))
• imgTVRD(/TRUE, /ORDER)
• window, 2
• tvscl, img, /true,/order

18
Writing to Raster files

• Many basic types of image formats can be written
  and read in IDL

19
Reading Writing 24-bit Images

• restore, 'test.data'
• device, decomposed0
• loadct,5
• SHADE_SURF, z, x, y, shadebytscl(dist(50))
• imgTVRD(/TRUE,/order)
• WRITE_TIFF, 'Cool.tif',img
• rread_tiff('Cool.tif')
• window, 2
• tvscl, r, /true,/order

TVRD returns a 3,n,m array img0,,R img1,,
G img2,,B
20
Reading Writing 8-bit Images

• restore, 'test.data'
• device, decomposed1
• loadct,5
• SHADE_SURF, z, x, y, shadebytscl(dist(50))
• imgTVRD(/order)
• tvlct, r, g, b, /get
• WRITE_TIFF, 'Cool.tif',img, redr, greeng,
  blueb
• write_bmp, 'Cool.bmp', reverse(img,2), r, g, b

TVRD returns a n,m array must read color table to
get associated colors
21
Making PS plots

• Postscript is a device independent plotting
  language
• Can be shown on screen or printers
• Can be contain both bitmap and vector graphics
• Vector graphics inherently better resolution
  because there are no pixels, just lines, circles,
  etc.
• Images are converted to pixels when the image is
  rendered, on whatever device is being used

22
Steps to make a PS plot

• First, set IDL to send graphics to a PS file
  using set_plot
• Then give IDL information about the PS file we
  will write using device
• Do our graphics calls
• All examples of plots, graphs, images,
  positioning, etc will work in PS files
• Close the PS file using the device command
• Set IDL back to the default graphics device (your
  screen)

23
Example PS Device

• restore, 'test.data' Get some data
• entry_device!d.name Record screen device name
• set_plot,'ps' Tell IDL we are using PS
• Next tell IDL about the PS file
• device, filename'test.ps', /inches, /color,
  xsize5, ysize5, xoffset1.75, yoffset3,
  bits8, /portrait, /helvetica
• loadct,5 Do our graphics stuff
• SHADE_SURF, z, x, y, shadebytscl(dist(50)),/font
• device,/close Close the PS file
• set_plot, entry_device Tell IDL to use screen
  again
• The file test.ps can now be printed, or viewed
  with a postscript viewer (photoshop, many others)

24
Example PS Device
25
Publication Quality Graphs

• When making images for JGR or a talk, it can
  often be hard to read default fonts, line sizes,
  character sizes, etc
• Try setting styles globally using
• !p.thick4 Globally set plot line thick 4
• !x.thick4 Globally set plot xaxis thick4
• !y.thick4
• Set plot commands with /font to use
• device, /helvetica, /bold, font_size16

26
Example

• decor2, /ps
• pro decor2, PSps
• if (keyword_set(PS)) then begin
• set_plot,'ps'
• device,/inches,xsize7.2,xoffset1
  ,ysize10, yoffset1,
• filenameoname'.ps',/helv
  etica,/bold,font_size16
• !P.thick4
• !x.thick4
• !y.thick4
• endif else begin
• window, 0, ysize1000, xsize800
• endelse

27
Example

• Plot commands here
• plot, x, y, /font
• if (keyword_set(PS)) then begin
• device,/close
• set_plot,'x'
• !P.thick1
• !x.thick1
• !y.thick1
• endif
• end

28
Example

• restore, 'test.data'
• entry_device!d.name
• set_plot,'ps'
• device, filename'test.ps', /inches, /color,
  xsize5, ysize5, xoffset1.75, yoffset3, bits8
  ,/helvetica,/bold,font_size16
• !P.thick4 !x.thick4 !y.thick4 !z.thick4
• loadct,5
• SHADE_SURF, z, x, y, shadebytscl(dist(50)),/font
• device,/close
• set_plot, entry_device

29
Result
30
The Z-Buffer!

• Like plotting on the screen, but instead you
  write the image into memory called the z-buffer
• Many images can be created stored in memory
  this way
• These images can rapidly be transferred to the
  screen to show a movie
• The images in the Z-buffer can be of any size
• Large composite images can be built in the
  z-buffer and then saved to a file, without using
  the screen
• The z-buffer is accessed like a PS file
• The z-buffer can be read using TVRD

What is the Z-Buffer?
31
Example Z-buffer

• restore, 'test.data' Get some data
• entry_device!d.name Record screen device name
• set_plot,'Z' Tell IDL we are using Z
• Next tell IDL about the Z
• device, set_resolution300,300
• loadct,5 Do our graphics stuff
• SHADE_SURF, z, x, y, shadebytscl(dist(50))
• rtvrd()
• device,/close Close the Z buffer
• set_plot, entry_device Tell IDL to use screen
  again
• tvscl, r

32
Example Z-buffer
33
Mapping

• IDL has the capability to draw maps using a
  number of cartographic projections
• Data can then be plotted on these maps in the
  current projection
• Unfortunately, routines are clumsy inaccurate
• Suggest using GMT instead
• Easier, better results
• Will show a few examples here

map_set, 0, 145 map_continents
34
Example Aitoff Data

• lon randomu(-100L, 100)360.0 - 180
• lat randomu(-200L, 100)180.0 -90
• map_set, /aitoff, /horizon
• map_continents
• oplot, lon, lat, psym1
• map_grid, /label

35
Mini-Homework

• Go to the web, download moderately sized image of
  your favorite planet (not Earth)
• Read it in
• Display it following the directions in book, 297
  298, in your favorite projection (but not
  mercator or cylindrical)

36
IDL Analysis Routines

• IDL has many analysis routines
• Appendix B in book
• Correlation
• Curve surface fitting
• Differentiation integration
• Eigenvalue/vector
• Gridding and interpolation
• Hypothesis testing
• Linear systems
• Error assessment

• Multivariate analysis
• Non-linear equations
• Optimization
• Probability
• Signal processing
• Sparse arrays
• Special functions
• Statistics
• Time series
• Transcendental functions
• Transforms

37
Least Squares Polynomial Fit

• To calculate a least squares fit to data, use
  poly_fit
• R POLY_FIT( X, Y, Degree, SIGMAvariable,
  YFITvariable)
• Note Returned variable R contains line fit
  parameters following
• yr0r1xr2x2rDeg-1xDeg
• !p.thick3
• xfindgen(10) yx
• plot, x, y
• y1x(randomu(-10L, 10)4-2)
• oplot, x, y1, line1
• rpoly_fit(x,y1,1, yfity2, sigmae)
• oplot, x, y2, line2

38
Least Squares Outliers

• To calculate a least squares fit to data, use
  poly_fit
• Result POLY_FIT( X, Y, Degree, SIGMAvariable,
  YFITvariable)
• xfindgen(10) yx
• plot, x, y, yrange0,18, /ystyle
• y1x(randomu(-10L, 10)4-2)
• y1815
• oplot, x, y1, line1
• rpoly_fit(x,y1,1, yfity2, sigmae)
• oplot, x, y2, line2

39
Min Abs Dev Fit Outliers

• More robust fits can be had with a minimum
  absolute deviation fit routine ladfit
• Result LADFIT( X, Y , ABSDEVvariable ,
  /DOUBLE )
• Note Returned variable R contains line fit
  parameters following
• yr0r1xr2x2
• xfindgen(10) yx
• y1x(randomu(-10L, 10)4-2)
• y1815
• plot, x, y1, yrange0,18, /ystyle
• rpoly_fit(x,y1,1, yfity2, sigmae)
• oplot, x, y2, line1
• rladfit(x,y1,absdeve)
• y3r0r1x
• oplot, x, y3, line2

40
Today

• Quiz 8 returned/Quiz 9 given
• Finish Graphics
• Shaded surface plots, large images
• Copying a screen image
• Reading and writing graphics files
• Other devices postscript and z-buffer
• Maps
• Homework

41
Next Time

• Maps and map projections
• Useful math functions in IDL
• Convolutions
• Correlations
• FFT's
• Wavelets
• Principle components