doc/localhost/ALE/download/ale-0.6.0-user/rendering/index.html

   1 <!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
   2 <html>
   3 <head>
   4   <title>ALE User Manual: Rendering</title>
   5
   6 <style type="text/css">
   7   TABLE.ba { max-width: 678; text-align: center; padding-bottom: 15; padding-top: 5}
   8   TABLE.inline { padding-right: 300; clear: left}
   9   TD.text_table {padding-left: 2; padding-right: 2; border-width: 1}
  10   H2 {clear: left}
  11   P {max-width: none; padding-right: 300; clear: left}
  12   BLOCKQUOTE {padding-right: 400 }
  13   LI {max-width: 640; clear: left}
  14   P.footer {max-width: none; width: auto; padding-left: 0}
  15   P.header {max-width: none; width: auto; padding-left: 0}
  16   HR.main {max-width: 640; clear: left; padding-left: 0; margin-left: 0}
  17   HR.footer {clear: both}
  18 </style>
  19 </head><body>
  20
  21
  22
  23 <table align=right valign=top width=160>
  24 <td valign=top height=600 width=160>
  25 <a href="http://auricle.dyndns.org/ALE/">
  26 <big>ALE</big>
  27 <br>
  28 Image Processing Software
  29 <br>
  30 <br>
  31 <small>Deblurring, Anti-aliasing, and Superresolution.</small></a>
  32 <br><br>
  33 <big>
  34 Local Operation
  35 </big>
  36 <hr>
  37 localhost<br>
  38 2684415918<br>
  39 </table>
  40
  41
  42
  43 <p><b>[ <a href="../">Up</a> ]</b></p>
  44 <h1>ALE Rendering</h1>
  45
  46 Rendering determines how an output image is constructed from a series of
  47 aligned input frames.  The incremental renderers, merging and drizzling, update
  48 the output image after reading each new frame from the series, whereas
  49 non-incremental renderers wait until the last frame has been loaded before
  50 producing output.  More than one renderer can be involved in producing the
  51 final output.  Available rendering options are outlined below.
  52
  53 <h2>Merging</h2>
  54
  55 Merging, the default incremental renderer, works most effectively on images
  56 that have no large changes in scale (e.g. camera zooming).  Since it uses
  57 bilinear interpolation, output images are approximately convolved with the
  58 triangle, or Bartlett, filter, and may appear to have reduced contrast and
  59 resolution when compared with the output of other renderers.
  60
  61 <h2>Drizzling</h2>
  62
  63 The drizzling renderer is an incremental renderer that can usually produce
  64 better images than merging, but it works most effectively with devices having
  65 very little blur and optical distortion.  Devices sampling very small angles of
  66 the scene at each pixel are ideal; hence, if very fine staircase aliasing is
  67 visible in the input frames, drizzling should work very well.  (For creation of
  68 higher resolution images with drizzling, use also the --scale option described
  69 below.)
  70
  71 <pre>
  72 --drizzle-diam=x  Drizzle with input pixel diameter x (where x &gt; 0)
  73 --drizzle-only    If drizzling, output black for pixels with no drizzle data.
  74 </pre>
  75
  76 <p>For more information about drizzling, see the paper by Richard Hook and
  77 Andrew Fruchter, "Variable-Pixel Linear Combination", or the following URL:
  78
  79 <p><a href="http://www.cv.nrao.edu/adass/adassVI/hookr.html">http://www.cv.nrao.edu/adass/adassVI/hookr.html</a>
  80
  81 <h2>Point-spread Function (PSF)</h2>
  82
  83 <p>Non-incremental renderers use a point-spread function to model the optical
  84 and electronic properties of the capture device used.  The PSF can be specified
  85 in terms of a linear and non-linear component.  The non-linear component is
  86 disabled by default.
  87
  88 <h3>Determining the PSF</h3>
  89
  90 <p>The point-spread function can be approximated through trial-and-error or
  91 based on known scenes.  ALE includes a script that numerically approximates the
  92 point-spread function based on images captured from known scenes.  This script
  93 can be found in the source distribution
  94 (<code>d2/render/psf/ale-psf-calibrate</code>).
  95
  96 <h3>Device Specification</h3>
  97
  98 If a device is known to ALE, then its PSF can be implicitly set with the device
  99 option.
 100
 101 <pre>
 102 --device &lt;d>      Set the capture device to &lt;d>.
 103 </pre>
 104
 105 Available devices are as follows:
 106
 107 <pre>
 108 xvp610_320x240    IBM PC Camera Pro (XVP610) in 320x240 mode.
 109 </pre>
 110
 111 <h3>PSF specification</h3>
 112
 113 <p>The PSF can also be explicitly specified by using one or both of the psf
 114 options.  These options have precedence over the device option.
 115
 116 <pre>
 117 --lpsf &lt;p>        Set linear colorspace point-spread function to &lt;p>
 118 --nlpsf &lt;p>       Set non-linear colorspace point-spread function to &lt;p>
 119 </pre>
 120
 121 Available explicit point-spread functions are as follows:
 122
 123 <pre>
 124 box=&lt;d>           Set the PSF to a box filter with diameter &lt;d>.
 125 stdin             Specify a custom filter at runtime.
 126 &lt;p>+&lt;p>           Summation of two functions
 127 </pre>
 128
 129 <h2>Unsharp Mask (USM) Renderer</h2>
 130
 131 <p>This non-incremental renderer uses an unsharp mask to compensate for
 132 blur introduced by the image capture device.  It is fairly inexpensive
 133 computationally, but it may not be effective in all cases.
 134
 135 <pre>
 136 --usm &lt;m>         Apply an unsharp mask with multiplier &lt;m>.
 137 </pre>
 138
 139 <h2>Irani-Peleg iterative solver</h2>
 140
 141 <p>This non-incremental renderer is very general, and attempts to reconstruct the
 142 original scene based on the captured frames and the PSF.  Unlike the USM
 143 renderer, it can compensate for blur introduced by the incremental
 144 renderers, and it may be more effective than the USM renderer in compensating
 145 for blur introduced by the image capture device.
 146
 147 <p>Note that this method can take a long time to produce output.  If
 148 incremental output is enabled (see below), output is written after each
 149 iteration is complete.  Using more iterations will generally improve output, up
 150 to the point of convergence, when additional iterations have no effect.  A dot
 151 is displayed to indicate the completion of each iteration.
 152
 153 <pre>
 154 --ips &lt;i>         Run &lt;i> iterations.
 155 </pre>
 156
 157 <p>For more information on the iterative backprojection technique of Michal Irani and
 158 Shmuel Peleg, see their paper, "Improving Resolution by Image Registration", at
 159
 160 <p><a href="ftp://ftp.wisdom.weizmann.ac.il/pub/irani/PAPERS/SR_CVGIP91.pdf">ftp://ftp.wisdom.weizmann.ac.il/pub/irani/PAPERS/SR_CVGIP91.pdf</a>
 161
 162 <h2>Exposure Registration</h2>
 163
 164 <p>By default, ALE adjusts an exposure model to minimize differences in tone
 165 between frames.  The option <code>--exp-noregister</code> indicates that
 166 exposure should be assumed uniform across all frames.  In this case, no
 167 adjustment is performed.</p>
 168
 169 <pre>
 170 --exp-register    Register exposure between frames.  [default]
 171 --exp-noregister  Assume uniform exposure across all frames.
 172 </pre>
 173
 174 For more information about analyzing and processing images having differing
 175 exposures, see Steve Mann's comparametric paper, at
 176
 177 <p><a href="http://wearcam.org/comparam.htm">http://wearcam.org/comparam.htm</a>
 178
 179 <h2>Certainty-weighted Rendering</h2>
 180
 181 <p>The exposure model in ALE includes an associated certainty function, which
 182 quantifies the reliability of sensor measurements.  By default, this function
 183 is raised to the zeroth power to obtain uniform certainty for all measurements.
 184 To specify a different certainty exponent, the option <code>--cx</code> can be
 185 used.
 186
 187 <pre>
 188 --cx &lt;x>          Render with certainty exponent &lt;x>
 189 --no-cx           Render with uniform certainty.  [default]
 190 </pre>
 191
 192 Note that ALE uses a one-sided certainty function, biased by
 193 backprojection value, in Irani-Peleg rendering.  In some cases, this approach
 194 seems to offer better results than the unbiased function used in drizzling and
 195 merging.  For more information on the certainty weighting approach used by ALE,
 196 see the ALE source code.
 197
 198 <p>For more information on Steve Mann's original approach to certainty-weighted
 199 rendering, see his comparametric paper, at
 200
 201 <p><a href="http://wearcam.org/comparam.htm">http://wearcam.org/comparam.htm</a>
 202
 203 <h2>Scale factor</h2>
 204
 205 Rendering with a larger scale factor causes internal data structures to store
 206 more information, resulting in image output of the specified larger scale.
 207 Larger scales can improve alignment precision and output image quality, but can
 208 also increase alignment and rendering times.
 209
 210 <pre>
 211 --scale=x         Scale images by the factor x (where x is at least 1.0)
 212 </pre>
 213
 214 <h2>Image extents</h2>
 215
 216 The --extend option records pixel data that falls outside of the region of the
 217 first frame in the sequence.  This can be useful for providing additional
 218 alignment constraints or creating panoramic image mosaics.  (See also the
 219 description of the --follow option in the <a href="../alignment">alignment</a>
 220 section.)
 221
 222 <pre>
 223 --extend          Increase image extents to accommodate all pixel data.
 224 --no-extend       Don't increase extents; crop to original frame. [default]
 225 </pre>
 226
 227 <h2>Incremental output</h2>
 228
 229 By default, the output file is written every time a renderer completes an
 230 update cycle.  In order to suppress output until the final image is rendered,
 231 specify the --no-inc option.
 232
 233 <pre>
 234 --inc             Produce incremental output.  [default]
 235 --no-inc          Don't produce incremental output.
 236 </pre>
 237
 238 <h2>Pixel replacement</h2>
 239
 240 <p>When using the merging or drizzling renderer, --replace indicates that each
 241 new frame should replace the corresponding region of the rendering, instead of
 242 being averaged with data from previous frames.  This option was originally
 243 suggested by Jeff Treece for use in stabilization of video streams.  His
 244 approach to stabilization is outlined below.
 245
 246 <h3>Video Stabilization</h3>
 247
 248 <p>If the files
 249
 250 <pre>0.ppm, 1.ppm, ..., 9.ppm</pre>
 251
 252 represent frames from an unstabilized video stream, then the following sequence
 253 of commands may result in a stabilized sequence:
 254
 255 <pre>
 256 ale --replace         0.ppm 1.ppm 1.stable.ppm
 257 ale --replace  1.stable.ppm 2.ppm 2.stable.ppm
 258 ale --replace  2.stable.ppm 3.ppm 3.stable.ppm
 259 .
 260 .
 261 .
 262 ale --replace  8.stable.ppm 9.ppm 9.stable.ppm
 263 </pre>
 264
 265 <p>The resulting stabilized sequence would be:
 266
 267 <pre>0.ppm, 1.stable.ppm, 2.stable.ppm, ..., 9.stable.ppm</pre>
 268
 269 <h3>Option Syntax</h3>
 270
 271 <pre>
 272 --replace         Replace overlapping areas when merging and drizzling.
 273 --no-replace      Do not replace.  [default]
 274 </pre>
 275
 276
 277 <br>
 278 <hr>
 279 <i>Copyright 2003, 2004 <a href="mailto:dhilvert@auricle.dyndns.org">David Hilvert</a></i>
 280 <p>Verbatim copying and distribution of this entire article is permitted in any medium, provided this notice is preserved.
 281
 282
 283 </body>
 284 </html>