Bug description
-AnotherRawTherapee exports TIFF files with '.tif'
-.tif or .TIF not visible with default filter; I have to switch to 'all' to find
-.tif or .TIF not processed with 'all files' filter; when selected, images are not loaded
-if I rename extensions to .tiff or .TIFF, the files are loaded and processed

Desktop (please complete the following information):

• OS:
  Linux Mint
• Version:
  app: 0.11.0
  os: 21.1
• Method of installation:
  APT repository on launchpad

Add a workflow to build on macos.

There might be some issues with compiler compatibility with some of the c++20 features, but in general it should work - all the dependencies are supported.

cv::imwrite doesn't support metadata writing, so the output files from Xpano currently have none

It would be nice to copy over some of the metadata from the input files + add a comment like "stitched from n panos by Xpano".

Probably will need to integrate https://github.com/Exiv2/exiv2

Add a bug report window accessible from the Help menu:

0. Put the BugReportPane object in PanoGui here, with implementation similar to the existing AboutPane.

1. Window contents

   • Display a message linking to https://github.com/krupkat/xpano/issues and my email for non-github users
   • Show the path to the log file directory, this is OS specific and is initialized here + here

2. There should be "copy to clipboard" buttons next to actionable items.

   • Use the imgui clipboard functions
   • They need to be initialized in main.cc like this (reddit)

3. Stretch goal: Parse the log file from the previous launch (xpano.1.log) and if there is an [error] entry show the bug report window indicating that there was an error in the last run.

   • Add a toggle "don't show this again", which will need to persist its state in the config

This issue will likely be split into multiple pull requests.

Bug description
Cloned latest release (today) and installed on OSX 13.1 with no errors and 48 tests pass.

Launching ~/xpano/install/bin/Xpano aborts due to dylib files not found.

fleming@Flemings-Mac-mini ~ % /Users/fleming/Documents/GitHub/xpano/install/bin/Xpano ; exit;
dyld[56853]: Library not loaded: @rpath/libopencv_imgcodecs.407.dylib
Referenced from: /Users/fleming/Documents/GitHub/xpano/install/bin/Xpano
Reason: tried: '/System/Volumes/Preboot/Cryptexes/OS@rpath/libopencv_imgcodecs.407.dylib' (no such file), '/usr/local/lib/libopencv_imgcodecs.407.dylib' (no such file), '/usr/lib/libopencv_imgcodecs.407.dylib' (no such file, not in dyld cache)
zsh: abort /Users/fleming/Documents/GitHub/xpano/install/bin/Xpano

Saving session...
...copying shared history...
...saving history...truncating history files...
...completed.

[Process completed]

These files are in ~/xpano/opencv/install/lib/cmake. Copying the aliases to /usr/local/lib did not fix issue.

I encountered this bug running the flatpak version on linux (gentoo profile amd64/17.1/desktop (stable))
Xpano cz.krupkat.Xpano 0.17.0 stable system

Looks like a dependancy is missing or an external call is broken ?

Thanks for this exciting bit of software! Can’t wait to try it.

Bug description

❯ ./misc/build/build-macos-12.sh
-- Performing Test NFD_USE_ALLOWEDCONTENTTYPES - Success
-- Found OpenCV: /Users/anileated/Developer/xpano/opencv/install (found version "4.7.0") found components: calib3d core features2d flann imgcodecs imgproc photo stitching 
CMake Error at CMakeLists.txt:87 (find_package):
  By not providing "Findspdlog.cmake" in CMAKE_MODULE_PATH this project has
  asked CMake to find a package configuration file provided by "spdlog", but
  CMake did not find one.

  Could not find a package configuration file provided by "spdlog" with any
  of the following names:

    spdlogConfig.cmake
    spdlog-config.cmake

  Add the installation prefix of "spdlog" to CMAKE_PREFIX_PATH or set
  "spdlog_DIR" to a directory containing one of the above files.  If "spdlog"
  provides a separate development package or SDK, be sure it has been
  installed.


-- Configuring incomplete, errors occurred!
See also "/Users/anileated/Developer/xpano/build/CMakeFiles/CMakeOutput.log".

❯ cmake --version
cmake version 3.24.4

Desktop (please complete the following information):

• OS: macOS 13.4
• Version: 41deb4b
• Method of installation: ./misc/build/build-macos-12.sh

Based on a comment by @hqhoang in #18 (comment)

Experiment with masking

• create gui for creating a mask for an input image
• get example images to test on
• experiment with customizing the masks argument in GraphCutSeamFinder called inside composePanorama

Initial situation
I imported 2 png files via the "Import images : Multiple files" button

• Their thumbnails display correctly in the bottom "film strip" style area
• In "list of pano" I have one pano titled "0,1" and a "show button
• Hovering the "show" button displays one more instance of the 2 thumbnails attach to my cursor position
• the right pano area is void

THE BUG :
Pushing the show buttons shows nothing
I have a "failed to stitch pano 0" with ERR_NEED_MORE_IMGS

ADDITIONAL INFO

• displaying debug info show a "matched" list with a functional "show" button that shows "184 inliers" displayed in green on the image

Desktop infos:

• OS: Gentoo - kernel 5.10.76-gentoo-r1 - profile 17.1 desktop (stable)
• Version: xpano 0.9.0
• Method of installation: flatpak

ONE TIME link to images
82.64.103.11/online/f.php?h=1OunIyVs
82.64.103.11/online/f.php?h=1OunIyVs

Base: Implement "Export all" action
Stretch goal: Implement "Export selected" action

Starting point:
Code location

Bug description

~ ls -al *.jpeg  | wc -l 
3075
~ Xpano * --output=/tmp/panorama.jpeg
...
[2023-09-30 17:36:50.131] [XPano] [error] Couldn't load images OpenCV(4.6.0) ./modules/flann/src/miniflann.cpp:521: error: (-215:Assertion failed) (size_t)knn <= index_->size() in function 'runKnnSearch_'

All the images are part of the same one panorama picture, shoot with dwarf 2 .

Desktop (please complete the following information):

• OS: Linux
• Version: 0.16.1
• Method of installation: dpkg

Window size should persist between launches.

This is already true for the individual ImGui windows thanks to its imgui.ini file.

The SDL window however is initialized with a constant size Code location.

Run clang-tidy check on push to main / pull requests.

Since clang-tidy requires a compilation database, it would make sense to slap this check on top of one of the existing build workflows, likely the Ubuntu one as it runs faster.

Existing windows script.

Hi,
Grag n dropping PNG files over the interface (whichever part) won't open files, even though the icon when hovering suggest it will.
Nothing is logged in the Displayable debug info (CTRL+D)

Having a large field of view (close to 180 degrees or more), doesn't play well with the perspective projection ("rectilinear", see wide field of view examples in https://www.cambridgeincolour.com/tutorials/image-projections.htm).

This can lead to very large output images, which can in turn eat up all memory trying to allocate space for it. As a first solution, let's stop the algorithm early and show the user that this is the case.

Howto

Replace the blender::prepare call here with the following:

auto result_roi = cv::detail::resultRoi(corners, sizes);
// check if result_roi.size() is too large and exit early
blender_->prepare(result_roi);

Add a new return type here:

• extend its ToString function.

Test on photos and make sure the app doesn't crash and shows the error.

Stretch goal

Make the max size of the panorama a user selectable value (in MPx):

• add the option here:

  xpano/xpano/algorithm/options.h

  Line 82 in dee2e45

  struct StitchUserOptions {

• Propagate the value to the Stitcher class here:

  xpano/xpano/algorithm/algorithm.cc

  Line 234 in dee2e45

  auto stitcher = stitcher::Stitcher::Create(cv::Stitcher::PANORAMA);

• Add a widget (probably an ImGui::InputInt) in the advanced menu here:

  xpano/xpano/gui/panels/sidebar.cc

  Line 308 in dee2e45

  if (debug_enabled) {

This will help solve #102

The license files are currently being loaded before the start of the main gui: Code location.

• Measure the impact / delay between launch and gui first showing up
• Implement async loading of the files

The error code from the OpenCV stitch function isn't being propagated to the gui.

Here is the call to the OpenCV function:
Code location

Propagate the value to the tooltip_message_ here:
Code location

We currently have the following supported extensions:

const std::array<std::string, 5> kSupportedExtensions = {"jpg", "jpeg", "tiff",
                                                         "png", "bmp"};

However on Linux this doesn't show files with uppercase extensions in the native file dialog.

Add preview during loading, so that the user has something else than the progress bar to watch :)

• Evaluate what kind of preview is worth the cost
• First step would probably be to show image previews as they're being loaded

E.g. by adding a cv::Mat LoadingPreview() function here.

Bug description
Opening a directory on Ubuntu 20.04 + flatpak crashes the app.

$ flatpak run cz.krupkat.Xpano 
terminate called after throwing an instance of 'std::filesystem::__cxx11::filesystem_error'
  what():  filesystem error: directory iterator cannot open directory: Not a directory [/run/user/1000/doc/f2ef324/image00.jpg]

Desktop

• OS: Ubuntu 20.04
• Version: 0.11.0
• Method of installation: flatpak

the multiblend option is missing even there're no errors in the build output

Desktop (please complete the following information):

• OS: macOS Ventura
• Version: 13.4
• arm64
• Method of installation: ./misc/build/build-macos-12.sh but in an macports environment

Add an option to export the source images for a pano.
This can be useful so that the user can stitch the images in a different app.

Starting point:
Code location

When I export a panoramic picture, the software reports an error:
ERR PANO TOO LARGE
Reset the adjustments through the edit menu.

Currently only subsequent images are being considered for grouping into panoramas.

Implement other strategies, e.g add an option to increase the considered neigborhood size.

Code location

building xpano in a macports environment (sdl2 spdlog catch2 installed via macports) fails:
output from console:
xpano_exiv0.28.0_fails.txt

it seems it doesn't use the included exiv2 lib but the system wide one installed via macports
so -Dexiv2_DIR=exiv2/install/lib/cmake/exiv2 doesn't seem to be sufficient

btw: compiles fine if macports exiv2 is deactivated ;)

I tried to generate a Panorama with 11 images, there I get the error
ERR_CAMERA_PARAMS_ADJUST_FAIL

Then I wanted to try another one with 8 images it seemed to work (It was stitched and adjusted correctly) until I wanted to export it, but it failed with:
ERR_PANO_TOO_LARGE

Images are 4160px width and 6240px height
I could provide them if needed.

With Hugin it was no problem to create the panorama from.

Please find below a tentative PKGBUILD for Arch users. Note that the ctest command in the check() section is commented out, until test #20 is fixed.

# Maintainer: Sébastien Guyader <[email protected]>

_pkgname=xpano
pkgname="$_pkgname-git"
pkgver=0.12.0.r1.g4d9afc6
pkgrel=1
pkgdesc="Automated photo stitching tool, with focus on simplicity and ease of use"
arch=('x86_64')
url='https://github.com/krupkat/xpano'
license=('GPL3')
depends=('gtk3' 'opencv>=4.4.0' 'sdl2' 'spdlog' 'catch2v3')
makedepends=('cmake' 'git')
provides=("$_pkgname")
conflicts=("$_pkgname")
source=('xpano::git+https://github.com/krupkat/xpano.git'
		'alpaca::git+https://github.com/p-ranav/alpaca.git'
		'imgui::git+https://github.com/krupkat/imgui.git'
		'nativefiledialog-extended::git+https://github.com/btzy/nativefiledialog-extended.git'
		'thread-pool::git+https://github.com/krupkat/thread-pool.git'
		)
md5sums=('SKIP' 'SKIP' 'SKIP' 'SKIP' 'SKIP')

pkgver(){
	cd $_pkgname
	git describe --long --tags --abbrev=7 | sed 's/^v//;s/\([^-]*-g\)/r\1/;s/-/./g'
}

prepare() {
	git -C xpano submodule init
	local _submodule
	for _submodule in alpaca imgui nativefiledialog-extended thread-pool
	do
		git -C xpano config --local submodule.external/${_submodule}.url "$srcdir/${_submodule}"
	done
	git -C xpano -c protocol.file.allow='always' submodule update
}

build() {
	cmake -B build -S $_pkgname \
		-DCMAKE_BUILD_TYPE=Release \
		-DCMAKE_INSTALL_PREFIX=/usr \
		-DBUILD_TESTING=ON \
		-DCMAKE_EXPORT_COMPILE_COMMANDS=OFF
	cmake --build build -j $(nproc)
}

check() {
	cd build
	#ctest
}

package() {
	DESTDIR="$pkgdir" cmake --install build
}

It doesn't look like Xpano is capable of seamless blending right now.

It becomes obvious when there's clear sky in the photos. Here is how hugin blends:

Here is what Xpano does:

You can clearly see darker patches of sky in the latter case.

There are multiple options how to deal with that. Enblend seems unmaintained, but https://horman.net/multiblend/ keeps getting updates apparently.

And then there's also https://github.com/weltyj/SkyFill to fill missing parts of the sky, if you are interested.

It would be great to use Xpano via the command line to enable scripting of panoramas.

This would also enable Xpano to be used as a darktable lua script. For example: https://github.com/darktable-org/lua-scripts/blob/master/contrib/hugin.lua allows you to select multiple images, which are then exported to a /tmp dir, imported and stitched by hugin, and the resulting panorama then is imported to darktable.

Bring in some of the options from stitching_detailed.

1. Identify which options would be useful
2. Implement

The goal is to keep the app simple to use and the defaults should "just work".

Calling the OpenCV function: Code location
Options menu: Code location

Currently Xpano converts inputs to 8bit images.

Stitching 16bit images with OpenCV was tested here: https://github.com/dankamongmen/stitcher16

todo: check how the current stitcher pipeline handles 16 bit inputs.

Bug description
If I try to stitch a panorama with a lot of (12, 15, 20 or more) images, the stitching progress is working normal from 0 up to a point of 90-99%.
But at this point, the progress slows down dramatically and memory usage is starting to increase, until all my RAM and swap is used and Xpano crashes (or get killed by the OS?).
Sometimes, this happens while calculating the preview, sometimes you can calculate the preview but it you calculate high res to export, it crashes.
I have 16GB RAM and 8GB swap, which is way more than a lot of other computer have.

Desktop (please complete the following information):

• OS: Ubuntu Linux 22.04
• Version: 0.15.1
• Method of installation: Flatpak

Bug description
Xpano crashes when minimizing it after loading images, video:
https://discuss.pixls.us/t/crash-when-minimizing-the-window/36538

Desktop (please complete the following information):

• OS: Windows 11
• Version: 0.11.0
• Method of installation: Any

The app by default loads images at half resolution and then computes the keypoints and descriptors.

These should be configurable in the options.

• resizing factor
• number of keypoints to detect per image
• keypoint/descriptor algorithms

Image loading: Code location
Loading dispatch: Code location
Options menu: Code location

Problem description
Currently the OpenCV stitch function cannot be cancelled from within Xpano.

This can lead to the progress appearing stuck (though the gui continues working).

When a user selects a new panorama / edits the current panorama by adding/removing an image a new stitch operation is enqueued.

The app behaves correctly in the sense that the new result is the one that will be shown to the user, but the old computation can still hog the CPU. In the worst case CPU can go to 100% and the app shows nothing for a long time.

Solution
Put "checkpoints" in the stitch function from where the function can be canceled.

The most unpredictable part of the pipeline is the estimateCameraParams function, so it would be worth to start there. The question to figure out is whether the changes needed will require using an OpenCV fork.

It would also be nice to figure out a way to get better progress report from this function (future issue probably).

Add an option to select registration resolution (in MPx) for the stitching algorithm. The images are downscaled to this size for finding matches between images.

The default value is 0.6MPx and that could be an issue e.g. if stitching two photos with electrical wires spanning over both of them.

Howto

• add a registration_resolution in:

  xpano/xpano/algorithm/options.h

  Line 82 in dee2e45

  struct StitchUserOptions {

• call SetRegistrationResolution in

  xpano/xpano/algorithm/algorithm.cc

  Line 232 in dee2e45

  StitchResult Stitch(const std::vector<cv::Mat>& images,

• Add a a widget in the gui for selecting the value (maybe ImGui::SliderFloat):

  xpano/xpano/gui/panels/sidebar.cc

  Line 301 in dee2e45

  Action DrawStitchOptionsMenu(pipeline::StitchAlgorithmOptions* stitch_options,

Stretch goal

Find / take some pictures where this effect can be demonstrated and add them to the PR.

After #54 added a bug report window, let's add automatic error log detection on startup:

• Parse the log file from the previous launch (xpano.1.log) and if there is an [error] entry show the bug report window indicating that there was an error in the last run.
• Add a toggle "don't show this again", which will need to persist its state in the config

In the workflow:

1. Run clang-format with the supplied .clang-format config file. (example script)
2. Return an error if git diff shows non zero diff.

Add support for stitching scans (https://docs.opencv.org/4.x/d2/d8d/classcv_1_1Stitcher.html#a114713924ec05a0309f4df7e918c0324ac706a6a118a4a648fef8b2fca5950e2a):

Howto

The value is selected here:

Extend stitching user options with a new enum (panorama / scans):

• the new enum will need a Label function and a constant array with all values, please check the other examples

Add a widget to select the value in gui (use utils::imgui::ComboBox for selecting an enum value):

If Scans mode is selected

• disable projection types widget + wave correction in the gui
• add an if statement in the Stitch function when setting up stitcher and set the values according to stitcher.cc

Stretch goal

Test on real scans, compare the two modes and add the images to the PR.

Implement export/import of the pano database.

• save urls to currently opened images and the panos that were detected by the program to a db file
• implement import of the db file, correctly handling missing images

Bug description
DE: XPano ist leider nur als riesiges FlatPak erhältlich. Wäre es möglch, es auch als AppImage anzubieten?
EN: XPano is unfortunately only available as huge FlatPak, Is it possible to offer it as well as Appimage?

Desktop (please complete the following information):

• OS: Debian Linux
• Version: LMDE 5 Elsie / Debian 11 „Bullseye“
• Method of installation: none yet