Refactor compare image

Summary

implements #2014 (closed) :

• UpdateLargestPossibleRegion() is not called in TestHelper::RegressionTestImage(...) anymore.
• Instead a Virtual writer is used to stream the DifferenceImageFilter, wich is now a persistent filter.
• Three tests have been added to test --compare-image :
  1. tsTvCompareImages_SameImages : test --compare-image with the same images. Example of output
  2. tsTvCompareImages_DifferentImages : test --compare-image with different images of same sizes (with ctest property WILL_FAIL set to true). Example of output
  3. tsTvCompareImages_DifferentSizes : test --compare-image with different images of different sizes (with ctest property WILL_FAIL set to true). Example of output
• The tests bfTuMeanShiftSmoothingImageFilterSpatialStability and otbGridResampleImageFilter have been modified to use the persistent version of DifferenceImageFilter
• The output images (baseline, test and diff) submitted to cdash are still png files, as tif cannot be displayed in the dashboard (only png and jpg are supported), and I don't know how to submit tif files anyway. Note that there is is a check on image size before generating the png.

Copyright

The copyright owner is CNES and has signed the ORFEO ToolBox Contributor License Agreement.

---

Check before merging:

• All discussions are resolved
• At least 2 👍 votes from core developers, no 👎 vote.
• The feature branch is (reasonably) up-to-date with the base branch
• Dashboard is green
• Copyright owner has signed the ORFEO ToolBox Contributor License Agreement
• Optionally, run git diff develop... -U0 --no-color | clang-format-diff.py -p1 -i on latest changes and commit