Support multiple distortion models, including a new division model

This commit makes it so CameraIntrinsics is no longer hardcoded
to use the traditional polynomial radial distortion model. Currently
the distortion code has generic logic which is shared between
different distortion models, but had no other models until now.

This moves everything specific to the polynomial radial distortion
to a subclass PolynomialDistortionCameraIntrinsics(), and adds a
new division distortion model suitable for cameras such as the
GoPro which have much stronger distortion due to their fisheye lens.

This also cleans up the internal API of CameraIntrinsics to make
it easier to understand and reduces old C-style code.

New distortion model is available in the Lens panel of MCE.

- Polynomial is the old well-known model
- Division is the new one which s intended to deal better with huge
  distortion.

Coefficients of this model works independent from each other
and for division model one probably want to have positive values
to have a barrel distortion.

1 files changed, 124 insertions, 83 deletions

diff --git a/extern/libmv/ChangeLog b/extern/libmv/ChangeLog
index 6ec5e0aa502..af18b02a4db 100644
--- a/extern/libmv/ChangeLog
+++ b/extern/libmv/ChangeLog
@@ -1,3 +1,127 @@
+commit ee21415a353396df67ef21e82adaffab2a8d2a0a
+Author: Sergey Sharybin <sergey.vfx@gmail.com>
+Date:   Thu Apr 17 16:26:12 2014 +0600
+
+    Support multiple distortion models, including a new division model
+    
+    This commit makes it so CameraIntrinsics is no longer hardcoded
+    to use the traditional polynomial radial distortion model. Currently
+    the distortion code has generic logic which is shared between
+    different distortion models, but had no other models until now.
+    
+    This moves everything specific to the polynomial radial distortion
+    to a subclass PolynomialDistortionCameraIntrinsics(), and adds a
+    new division distortion model suitable for cameras such as the
+    GoPro which have much stronger distortion due to their fisheye lens.
+    
+    This also cleans up the internal API of CameraIntrinsics to make
+    it easier to understand and reduces old C-style code.
+    
+    Reviewers: keir
+    
+    Reviewed By: keir
+    
+    CC: jta
+    
+    Differential Revision: https://developer.blender.org/D335
+
+commit 313252083f6dfa69a93c287bed81dec616503c1b
+Author: Sergey Sharybin <sergey.vfx@gmail.com>
+Date:   Tue Apr 15 18:23:38 2014 +0600
+
+    Fix failure of the image transform linear test
+    
+    Mainly was caused by the flakyness of image rotation in cases
+    when image has even size. The test was expecting the transform
+    code to rotate the image around pixel corner, which isn't a
+    common behavior in image processing applications. Rotation
+    is usually done around the pixel center.
+    
+    So now made it so RotateImage() rotates the image around the
+    pixel center which gives 100% proper result for odd sized images
+    (i.e. center pixel stays untouched).
+    
+    Also made the tests to use odd image sizes which are more
+    predictable by the humans. We can use even sized images in the
+    tests as well but their result wouldn't be so much spectacular.
+    
+    Another issue with the tests was caused by RescaleImageTranslation
+    test which did expect things which are not happening in the
+    function.
+    
+    Reviewers: keir
+    
+    Reviewed By: keir
+    
+    Differential Revision: https://developer.blender.org/D463
+
+commit 80d6945bf5f996b97cd41df0e422afce5e10e7f9
+Author: Sergey Sharybin <sergey.vfx@gmail.com>
+Date:   Mon Apr 14 00:01:32 2014 +0600
+
+    Unit tests for feature detector
+    
+    Currently covers only simplest cases with synthetic images.
+    Also at this point mainly Harris detector is being testes,
+    other detectors behaves a bit unexpected on synthetic images
+    and this is to be investigated further.
+    
+    Tests will be extended further later.
+    
+    Additional change:
+    
+    - Added constructor to Feature structure
+    - Added operator << for feature for easier debug dumps.
+    
+    TODO: Some tests are not giving the result which i was expected
+    to. This is to be investigated further by finding the reference
+    detector implementation. For until then keeping that tests
+    commented out.
+    
+    Reviewers: keir
+    
+    Reviewed By: keir
+    
+    Differential Revision: https://developer.blender.org/D316
+
+commit 397c3d3ed46eb4967eb285c8369cc125bea4b132
+Author: Sergey Sharybin <sergey.vfx@gmail.com>
+Date:   Fri Apr 4 16:17:57 2014 +0600
+
+    Compilation error fix
+    
+    Not totally sure why this is needed, but multiview indeed
+    uses V3D library still, so it needs to be linked against it.
+    
+    Patc by Martijn Berger, thanks!
+
+commit 1c36279239cbffe152493106eb04e55df7ebd649
+Author: Sergey Sharybin <sergey.vfx@gmail.com>
+Date:   Fri Apr 4 14:03:43 2014 +0600
+
+    Upgrade Eigen to 3.2.1 version
+    
+    To main reasons for this:
+    - Probably this would solve strict compiler warnings
+    - It brings new stuff like sparse LU decomposition which
+      might be useful in the future.
+
+commit de698f442934f475478463445f78a00ea632e823
+Author: Sergey Sharybin <sergey.vfx@gmail.com>
+Date:   Thu Apr 3 15:08:26 2014 +0600
+
+    Fix compilation error when using make from the sources root
+    
+    - Don't force flann to be static. It's a general rule on linux
+      to have dynamic libraries for all the bits instead of having
+      statically-linked dynamic libraries.
+    
+    - Some weirdo stuff was happening around OpenExif, it was only
+      built on Apple, so don't link targets against this lib on
+      other platforms.
+    
+    - Some libraries were missing for qt-tracker.
+
 commit 901b146f28825d3e05f4157ca2a34ae00261b91a
 Author: Sergey Sharybin <sergey.vfx@gmail.com>
 Date:   Wed Mar 26 17:44:09 2014 +0600
@@ -550,86 +674,3 @@ Date:   Sun May 12 22:34:54 2013 +0600
     - Removed note about unsupported camera intrinsics
       refining flags. Technically, all combinations
       are possible.
-
-commit 4432eb80f27e929f8750229aaada625d4f3ac5ee
-Author: Sergey Sharybin <sergey.vfx@gmail.com>
-Date:   Sun May 12 22:19:31 2013 +0600
-
-    Remove check for zero focal length in BA cost functor
-    
-    This check is actually redundant, because empty intrinsics
-    will have focal length of 1.0, which means original comment
-    about BundleIntrinsics was not truth.
-    
-    It is possible that external user will send focal length of
-    zero to be refined, but blender prevents this from happening.
-
-commit 34a91c9b8acb0dba3382866fbd29bb9884edb98a
-Author: Sergey Sharybin <sergey.vfx@gmail.com>
-Date:   Sat May 11 20:33:54 2013 +0600
-
-    Fix compilation error with msvc2012
-    
-    Using change from glog's upstream for this.
-
-commit 87be4f030d025e4b29d9243d12bc458b2bb6762a
-Author: Sergey Sharybin <sergey.vfx@gmail.com>
-Date:   Sat May 11 19:50:57 2013 +0600
-
-    Style cleanup, mainly pointed by Sameer in Ceres's codereview
-
-commit 7fa9c0b83d5e0fbd331add2952045076c2028d1b
-Author: Sergey Sharybin <sergey.vfx@gmail.com>
-Date:   Fri May 10 18:30:40 2013 +0600
-
-    Keyframe selection improvements
-    
-    Added additional criteria, which ignores
-    keyframe pair if success intersection factor
-    is lower than current candidate pair factor.
-    
-    This solves issue with keyframe pair at which
-    most of the tracks are intersecting behind the
-    camera is accepted (because variance in this
-    case is really small),
-    
-    Also tweaked generalized inverse function,
-    which now doesn't scale epsilon by maximal
-    matrix element. This gave issues at really bad
-    candidates with unstable covariance. In this
-    case almost all eigen values getting zeroed
-    on inverse leading to small expected error.
-
-commit 0477ef1aa8fc92848f03c45e32539210be583b80
-Author: Sergey Sharybin <sergey.vfx@gmail.com>
-Date:   Fri May 10 17:59:40 2013 +0600
-
-    Keyframe selection code cleanup
-    
-    - Updated comments in code.
-    - Removed currently unused functions.
-      Actually, they'd better be moved to some generic
-      logging module, but we don't have it now so was
-      lazy to create one now. Could happen later using
-      code from git history
-    - Renamed function to match better to what's going
-      on in it.
-
-commit fee2d7cc6003942f628c9a24b74008fd491b85b9
-Author: Sergey Sharybin <sergey.vfx@gmail.com>
-Date:   Fri May 10 17:44:49 2013 +0600
-
-    Optimization for reconstruction variance
-    
-    Achieved by replacing SVD-based pseudo-inverse with
-    an eigen solver pseudo inverse.
-    
-    New function works in the same way: it decomposes
-    matrix to V*D*V^-1, then inverts diagonal of D
-    and composes matrix back.
-    
-    The same way is used to deal with gauges - last
-    7 eigen values are getting zeroed.
-    
-    In own tests gives approx 2x boost, without
-    visible affect on selected keyframe quality.