mypapers/drafts/geg2210/assignment-8.html

   1 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
   2 <html>
   3   <head>
   4     <title></title>
   5   </head>
   6
   7   <body>
   8     <h1></h1>
   9
  10 Here are my notes from today.
  11
  12
  13 Logged into jern.uio.no using ssh to run ERDAS Imagine.  Started by
  14 using 'imagine' on the command line.  The images were loaded from
  15 /mn/geofag/gggruppe-data/geomatikk/
  16
  17 Tried to use svalbard/tm87.img, but it only have 5 bands.  Next tried
  18 jotunheimen/tm.img, which had 7 bands.
  19
  20 The pixel values in a given band is only a using a given range of
  21 values.  This is because sensor data in a single image rarely extend
  22 over the entire range of possible values.
  23
  24 Evaluation of the different bands
  25 =================================
  26
  27 band 1, blue (0.45-0.52 um)
  28 ---------------------------
  29
  30   Visible light, and will display a broad range of values both over
  31   land and water.  Reflected from ice, as those are visible white and
  32   reflect all visible light waves.  Histogram show most values between
  33   30 and 136.  Mean values of 66.0668.  There are one wide peak with
  34   center around 50.  There are two peaks at 0 and 255.
  35
  36 band 2, green (0.52-0.60 um)
  37 ----------------------------
  38
  39   Visible light, and will display a broad range of values both over
  40   land and water.  Reflected from ice, as those are visible white and
  41   reflect all visible light waves.  Histogram show most values from 8
  42   to 120.  The mean value is 30.9774.  There are two main peaks at 20
  43   and 27.  There is also a pie at 0.
  44
  45 band 3, red (0.60-0.69 um)
  46 --------------------------
  47
  48   Visible light, and will display a broad range of values both over
  49   land and water.  Reflected from ice, as those are visible white and
  50   reflect all visible light waves.  Histogram show most values from 33
  51   t 135, with one wide peak around 52.  There are also seem to be two
  52   peaks at 0 and 255.  The mean value is 34.3403.
  53
  54 band 4, near-infraread (0.76-0.90 um)
  55 -------------------------------------
  56
  57   Water acts as an absorbing body so in the near infrared spectrum,
  58   water features will appear dark or black meaning that all near
  59   infrared bands are absorbed. On the other hand, land features
  60   including ice, act as reflector bodies in this band.  The histogram
  61   show most values between 7 and 110.  The mean is 40.1144.  There are
  62   two peaks at 7 and 40.
  63
  64 band 5, mid-infrared (1.55-1.75 um)
  65 -----------------------------------
  66
  67   The ice, glaciers and water do not reflect any mid-infrared light.
  68   The histogram show most values between 1 and 178.  The mean is
  69   49.8098 and there are two peaks at 6 and 78, in addition to two
  70   peaks at 0 and 255.
  71
  72 band 6, thermal infrared (10.4-12.5 um)
  73 ---------------------------------------
  74
  75   Display the temperature on earth.  We can for example see that the
  76   ice is colder than the surrounding areas.  The histogram show most
  77   values between 36 to 122.  The mean is 102.734.  There are one wide
  78   peak around 53, in addition to two peaks at 0 and 255.
  79
  80 band 7, mid-infrared (2.08-2.35 um)
  81 ------------------------------------
  82
  83   The ice, glaciers and water do not reflect any mid-infrared
  84   frequencies.  The histogram show most values between 77 and 150.
  85   The mean is 24.04, and there are one wide peak at 130 and a smaller
  86   peak at 83, in addition to one peak at 0.
  87
  88 Image enhancement
  89 -----------------
  90
  91 We can get a good contrast stretch by using the histogram
  92 equalisation.  This will give us the widest range of visible
  93 separation between features.
  94
  95 Displaying colour images
  96 ------------------------
  97
  98 Comparing a map we found on the web,
  99 <URL:http://home.online.no/~oe-aase/jotunheimen/jotun2000topper.jpg.>
 100 and the standard infrared image composition, we can identify some
 101 features from the colors used:
 102
 103  - water is black or green
 104
 105  - ice and glaciers are white, while snow is light green.
 106
 107  - vegetation is red.
 108
 109  - non-vegetation is brown or dull red when closer to snow and
 110    glaciers.
 111
 112 Next, we tried to shift the frequencies displayed to use blue for the
 113 red band, green for the near ir band and red for the mid ir (1.55-1.75
 114 um).  With this composition, we get some changes in the colours of
 115 different features:
 116
 117  - water is black
 118
 119  - ice and glaciers are light blue, while snow is dark blue.
 120
 121  - vegetation is light green and yellow.
 122
 123  - non-vegetation is red or brown.
 124
 125 Filtering and image sharpening
 126 ==============================
 127
 128 We decided to work on the grey scale version of the thermal infrared.
 129 This one has lower resolution then the rest of the bands, with 120m
 130 spatial resolution while the others have 30m spatial resolution.
 131
 132 The high pass filtering seem to enhance the borders between the
 133 pixels.  Edge detection gave us the positions of glaciers and water.
 134 We tried a gradient filter using this 3x3 matrix: [ 1 2 -1 / 2 0 -2 /
 135 1 -2 -1 ].  It gave a similar result to the edge detection.
 136
 137
 138 We also tried unsharp filtering using this 3x3 matrix: [ -1 -1 -1 / -1
 139 8 -1 / -1 -1 -1 ].  This gave similar results to the edge detection
 140 too.
 141
 142 We started to suspect that the reason the 3x3 filters gave almost the
 143 same result was that the fact that the spatial resolution of the
 144 thermal band is actually 4x4 pixels.  Because of this, we tried with a
 145 5x5 matrix, making sure it sums up to 0.
 146
 147    -1 -1 -1 -1 -1
 148    -1 -1 -1 -1 -1
 149    -1 -1 24 -1 -1
 150    -1 -1 -1 -1 -1
 151    -1 -1 -1 -1 -1
 152
 153 Next, we tried some different weight:
 154
 155    -1 -1 -1 -1 -1
 156    -1 -2 -2 -2 -1
 157    -1 -2 32 -2 -1
 158    -1 -2 -2 -2 -1
 159    -1 -1 -1 -1 -1
 160
 161 This one gave more lines showing the borders between the thermal
 162 pixels.
 163
 164 From: shanette Dallyn <shanette_dallyn@yahoo.ca>
 165 Subject: Re: My notes from todays exercise
 166 To: Petter Reinholdtsen <pere@hungry.com>
 167 Date: Sat, 30 Apr 2005 15:16:59 -0400 (EDT)
 168
 169 Hey Petter!
 170  Allright, I looked up some stuff on statistics and the most valuable
 171 conclusion that I can come up with for the histograpm peak question is:
 172 "The peak values of the histograms represent the the spectral sensitivity
 173 values that occure the most often with in the image band being analysed"
 174
 175 For the grey level question go to http://www.cs.uu.nl/wais/html/na-dir/sci/
 176 Satellite-Imagery-FAQ/part3.html I found this and thought that the first major
 177 paragraph pretty much answered the question for the grey levels.
 178
 179 theory of convolution:
 180
 181                       Specialty Definition: Convolution
 182
 183                    (From Wikipedia, the free Encyclopedia)
 184
 185 In mathematics and in particular, functional analysis, the convolution
 186 (German: Faltung) is a mathematical operator which takes two functions and and
 187 produces a third function that in a sense represents the amount of overlap
 188 between and a reversed and translated version of .
 189
 190 The convolution of and is written . It is defined as the integral of the
 191 product of the two functions after one is reversed and shifted.
 192
 193 The integration range depends on the domain on which the functions are
 194 defined. In case of a finite integration range, and are often considered as
 195 cyclically extended so that the term does not imply a range violation. Of
 196 course, extension with zeros is also possible.
 197
 198 If and are two independent random variables with probability densities and ,
 199 respectively, then the probability density of the sum is given by the
 200 convolution .
 201
 202 For discrete functions, one can use a discrete version of the convolution. It
 203 is then given by
 204
 205 When multiplying two polynomials, the coefficients of the product are given by
 206 the convolution of the original coefficient sequences, in this sense (using
 207 extension with zeros as mentioned above).
 208
 209 Generalizing the above cases, the convolution can be defined for any two
 210 square-integrable functions defined on a locally compact topological group. A
 211 different generalization is the convolution of distributions.
 212
 213 I hope this will help!
 214
 215 Shanette
 216
 217     <hr>
 218     <address><a href="mailto:pere@hungry.com">Petter Reinholdtsen</a></address>
 219 <!-- Created: Sun May  1 13:25:38 CEST 2005 -->
 220 <!-- hhmts start -->
 221 Last modified: Sun May  1 13:32:31 CEST 2005
 222 <!-- hhmts end -->
 223   </body>
 224 </html>