For algorithmic changes related to the photogrammetric pipeline, please refer to AliceVision changelog.
Based on AliceVision 3.0.0.
- Major improvements of the depth map quality, performances and scalability. The full resolution can now be computed on most of the standard GPUs.
- FeatureExtraction is now using DSP-SIFT by default for the 3D Reconstruction pipeline.
- Capacity to create panoramas with very high resolutions using a limited amount of memory.
- Enhanced interpretation of RAW images, including new support for Adobe Digital Camera Profile and Lens Camera Profiles databases (if installed on your workstation).
- Improved color management with OCIO support and more options to export in various colorspaces including ACEScg.
- New graph templates enabling users to create custom pipelines.
- Expose a new experimental pipeline for Camera Tracking.
- Improved GraphEditor with copy-paste and multi-selection.
- Improved ImageGallery with thumbnails cache and search options.
- 2D Viewer is now using floating-point images by default.
- And a very large amount of UI improvements and bug fixes.
- [nodes] DepthMap: depth map improvements PR
- Integration of AprilTag library according to issue #1179 and AliceVision pull request #950 PR
- [nodes] add gps option to SfMTransform PR
- [ui] add support for selecting multiple nodes at once PR
- Image Gallery: Add a menu to set the StructureFromMotion initial pair from the gallery PR
- Texturing Color Space PR
- Add support for Lens Camera Profiles (LCP) PR
- RAW advanced processing PR
- Add new file watcher behaviours PR
- Add internal attributes in "Notes" tab PR
- New nodes for large memory use in panoramas PR
- [ui] Thumbnail cache PR
- [nodes] new SfMTriangulation node PR
- Color management for RAW images PR
- [ui] image gallery search bar PR
- [ui] Viewer 2D: enable the HDR viewer by default PR
- [ui] Improve the manipulator of the panorama viewer PR
- Color space management PR
- Show generated images in 2D viewer when double-clicking on node PR
- [ui] Elapsed time indicators in log PR
- [nodes] SfMTransform: add auto_from_cameras_x_axis PR
- Graph Editor: Support copy/paste of selected nodes and scene import PR
- [Feature Matching] Add an option to remove matches without enough motion PR
- Output in ACES or ACEScg color space PR
- Use project files to define pipelines PR
- [nodes] StructureFromMotion: Add option computeStructureColor PR
- [core] add env var to load nodes from multiple folders PR
- Depth map refactoring PR
- Draft Reconstruction pipeline PR
- [ui] Add filters to image gallery PR
- [nodes] New node "RenderAnimatedCamera" using blender API PR
- New node to import known poses for various file formats PR
- New ImageMasking and MeshMasking nodes PR
- Create Split360Images Node PR
- New lens distortion calibration node PR
- New experimental camera tracking pipeline PR
- [multiview] New pipeline "Photogrammetry and Camera Tracking" PR
- [nodes] KeyframeSelection: Rework the node and add parameters for new selection methods PR
- [nodes] ImageProcessing: Add and hide the fringing correction in the LCP PR
- Update highlight mode description in imageProcessing node PR
- [ui] Prompt a warning dialog when attempting to submit an unsaved project PR
- [panorama] force pyramid levels count in compositing PR
- [ui] Add a new advanced menu action to load templates like regular projects PR
- [panorama] New option to disable compositing tiling PR
- [sfmtransform] Transformation parameter availability PR
- Apply DCP metadata in imageProcessing PR
- [ui] FeaturesViewer: track endpoints PR
- LdrToHdrMerge node: Add a checkbox enabling the manual setting of the reference bracket for HDR merging PR
- [ui] Display nodes computed in another Meshroom instance as "Computed Externally" PR
- [ui] Use the location of the most recently imported images as the base folder for the "Import Images" dialog PR
- [ui] GraphEditor: use maxZoom to fit on nodes PR
- [ui] Viewer2D: support all Exif orientation tags PR
- Use DCP by default if the database is set and create errors on missing DCP files PR
- [ui] Load 3D Depth Map: minor improvements PR
- [ui] Checkbox to enable/disable 8-bit viewer PR
- Add Ripple submitter PR
- [ui] ImageGallery: Increase the GridView's cache capacity PR
- [ui] Reorganize the "File" menu PR
- [nodes] rename: remove "utils" from executables names PR
- [ui] Integrate QtOIIO into QtAliceVision PR
- Add nl means denoising open cv in image processing node PR
- [core] Add cgroups support to meshroom PR
- Remove support for Python 2 PR
- [submitters] Add an option to update the job title on submitters PR
- [ui] GraphEditor: create new pipelines with the node menu PR
- [bin] meshroom_batch: allow passing list of values to param overrides PR
- [ui] ImageGallery: update the Viewer2D correctly when the GridView's current item changes PR
- [ui] keyboard shortcut: press tab to open node menu PR
- Update bounding box display to use the correct geometric frame PR
- [ui] Paste nodes at the center of the Graph Editor when it does not contain the mouse PR
- Use most recent project as base folder for file dialogs PR
- [ui] Restrain the "copy/paste nodes" shortcuts to the GraphEditor PR
- [core] Set the "template" flag to "false" when saving a project as a regular file PR
- [ui] Display computation time for "running" or "finished" nodes PR
- Removed duplicated call to findnodes PR
- Add dedicated "minimal" mode for templates PR
- [ui] Reduce confusion when qml loading fails PR
- [ui] Update intrinsics table when switching between groups PR
- [bin] batch: allow to set params inside groups PR
- [camerainit] update parameters to use focal in mm PR
- [bin] newNodeType: update PR
- [minor] renderfarm submission with rez PR
- [ui] widgets visibility options PR
- [bin] Avoid multi-threading in non-interactive computation PR
- [nodes] Mesh*: use file extension to choose the file format PR
- Upgrade Texturing node and add multiples mesh file types PR
- Optical center relative to the image center PR
- [core] Improve project files upgrade PR
- [ui] Add a clear images button PR
- [ui] highlight the edge that will be deleted PR
- Update 2d viewer for new Track drawing mode of QtAliceVision PR
- Add cli script to start Meshroom on Windows PR
- Allow replacing edges PR
- No cmd line range arguments if we have only a single chunk PR
- [nodes] ExportAnimatedCameras: new sfmDataFilter parameter PR
- Node highlight radius PR
- [ui] Fix conditions on which the prompt asking the user to save a project before submitting it to the render farm relies PR
- [ui] ImageGallery: Allow image drop if the active group is not computing PR
- [ui] Viewer2D: fix displayed metadata PR
- [setup] add all scripts in bin/ as executables PR
- Add a unit test to check the node versions of templates PR
- [nodes] Split360Images: update attributes to software version 2.0 PR
- [ci] upgrade github actions rules PR
- Update INSTALL.md PR
- [docs] Python documentation generation using Sphinx PR
- Documentation update : how to use Meshroom without building AliceVision PR
- [pipelines] Panorama: Fix inputs of the "Publish" nodes PR
- [nodes] ExportAnimatedCameras: fix output params labels PR
- [nodes] PanoramaWarping: remove obsolete image output attributes PR
- Fix the documentation related to Panorama nodes PR
- Fix missing Publish nodes in templates PR
- [ui] Intrinsics: Fix warnings and exceptions PR
- [ui] fix thumbnail cache bugs PR
- [ImageGallery] Match the filter selection with the gallery's display PR
- [ui] fix "Sync Camera with Image Selection" PR
- Fix exceptions raised when accessing attributes that either do not exist or are not associated to a graph PR
- fix(sec): upgrade psutil to 5.6.7 PR
- [ui] Fix all "TypeError" QML warnings PR
- [ui] Viewer2D: fix minor issues PR
- Fix crash when importing images with non-ascii characters in their filepath PR
- Fix and prevent mismatches between an attribute's type and its default value's type PR
- Fix various typos PR
- [ui] ImageGallery: fix some minor issues PR
- [core] fix logging of nodes loading PR
- Fix node duplication/removal behaviour PR
- [ui] Fix offset between the mouse's position and the tip of the edge when connecting two nodes PR
- Fix compatibility with Python 3 PR
- Fix stats PR
- [ui] ImageGallery: fix missing function changeCurrentIndex PR
- [UI] StatViewer: fix displayed unit PR
- [ui] fix uvCenterOffset PR
- Fix meshroom_batch PR
- Fix incompatibility with recent cx_Freeze PR
- [bin] meshroom_batch: fix typo in pipeline names PR
- Removing
io_counters
from the ProcStatatistics PR - Fix NameError PR
- [ui] Image Gallery: Fix the display of the intrinsics table with temporary CameraInit nodes PR
- [ui] Correctly update the Viewer 2D when there are temporary CameraInit nodes PR
- [ui] Clear Images: Request a graph update after resetting the viewpoints and intrinsics PR
- [ui] Improve "Clear Images" action's behaviour and performance PR
- [Viewer] Load and unload the SfMStats components explicitly every time they are shown and hidden PR
- [ui] Drag&Drop: Use a pool of threads for asynchronous intrinsics computations PR
- [nodes] CameraInit: upgrade version following the parameters changes PR
- [ui] app: temporary workaround for qInstallMessageHandler PR
- [ui] ImageGallery: fix the DB path in the "Edit Sensor Database" dialog PR
- [ui] Correctly determine if a graph is being computed locally and update nodes' statuses accordingly PR
- [nodes] CameraInit: all intrinsics parameters should invalidate PR
- [ci] add bug to the list of tag to skip the stale check PR
- Fix various typos in the source code PR
- Update ion startup PR
- New script to launch meshroom under ion environment PR
- [doc] fix the bibtex PR
- [doc] readme: add citation PR
Thanks to Fabien Servant, Gregoire De Lillo, Vincent Demoulin, Thomas Zorroche, Povilas Kanapickas, Simone Gasparini, Candice Bentejac, Loic Vital, Charles Johnson, Jean Melou, Matthieu Hog, Simon Schuette, Ludwig Chieng, Vincent Scavinner, Nils Landrodie, Stella Tan for the major contributions.
Other release contributors: asoftbird, DanielDelaporus, DataBeaver, elektrokokke, fabiencastan, Garoli, ghost, hammady, luzpaz, MakersF, pen4, remmel, wolfgangp
Based on AliceVision 2.4.0.
- [panorama] PanoramaCompositing: new algorithm with tiles to deal with large panoramas PR
- [feature] Improve robustness of sift features extraction on challenging images: update default values, add new filtering and add dsp-sift variation PR
- [ui] Improve Graph Editor UX with better visualization of nodes connections, the ability to accumulate nodes to compute locally or the ability to compute multiple branches in parallel on renderfarm with a new locking system per node, etc. PR
- [nodes] Meshing: improve mesh quality with a new post-processing. Cells empty/full status are filtered by solid angle ratio to favor smoothness. PR
- [nodes] MeshFiltering: smoothing & filtering on subset of the geometry PR
- [ui] Viewer: fix gain/gamma behavior and use non-linear sliders PR
- [core] taskManager: downgrade status per chunk PR
- [core] Improve graph dependencies: dependencies to an input parameter is not a real dependency PR
- [nodes] Meshing: Add
addMaskHelperPoints
option PR - [nodes] Meshing: More control on graph cut post processing PR
- [nodes] Meshing: new cells filtering by solid angle ratio PR
- [nodes] Meshing: add seed and voteFilteringForWeaklySupportedSurfaces PR
- [nodes] Add some mesh utilities nodes PR
- [nodes] SfmTransform: new from_center_camera PR
- [nodes] Panorama: new options to init with known poses PR
- [nodes] FeatureMatching: add cross verification PR
- [nodes] ExportAnimatedCamera: New option to export undistort maps in EXR format PR
- [nodes] new wip node
LightingEstimation
to estimate spherical harmonics from normal map and albedo PR - [nodes] CameraInit: add a boolean for white balance use PR
- [ui] fix error on live reconstruction PR
- [ui] init saveAs folder PR
- [ui] add link to online documentation in 'Help' menu PR
- [ui] New node menu categories PR
Based on AliceVision 2.3.1.
- [core] Fix crashes on process statistics (windows-only) PR
Based on AliceVision 2.3.0.
- [nodes] New Panorama Stitching nodes with support for fisheye lenses PR PR
- [nodes] HDR: Largely improved HDR calibration, including new LdrToHdrSampling for optimal sample selection PR PR PR
- [ui] Viewer3D: Input bounding box (Meshing) & manual transformation (SfMTransform) thanks to a new 3D Gizmo PR
- [ui] Sync 3D camera with image selection PR
- [ui] New HDR (floating point) Image Viewer PR
- [ui] Ability to load depth maps into 2D and 3D Viewers PR PR
- [ui] New features overlay in Viewer2D allows to display tracks and landmarks PR PR
- [ui] Add SfM statistics PR
- [ui] Visual interface for node resources usage PR
- [nodes] Coordinate system alignment to specific markers or between scenes PR
- [nodes] New Sketchfab upload node PR
- [ui] Dynamic Parameters: add a new 'enabled' property to node's attributes PR PR
- [ui] Viewer: add Camera Response Function display PR PR
- [ui] UI improvements in the Viewer2D and ImageGallery PR
- [bin] Improve Meshroom command line PR PR
- [nodes] New ImageProcessing node PR PR PR
- [nodes]
FeatureMatching
Addfundamental_with_distortion
option PR - [multiview] Declare more recognized image file extensions PR
- [multiview] More generic metadata support PR
- [nodes] CameraInit: New viewId generation and selection of allowed intrinsics PR
- [core] Avoid error during project load on border cases PR
- [core] Compatibility : Improve list of groups update PR
- [core] Invalidation hooks PR
- [core] Log manager for Python based nodes PR
- [core] new Node Update Hooks mechanism PR
- [core] Option to make chunks optional PR
- [nodes] Add methods in ImageMatching and features in StructureFromMotion and FeatureMatching PR
- [nodes] FeatureExtraction: add maxThreads argument PR
- [nodes] Fix python nodes being blocked by log PR
- [nodes] ImageProcessing: add new option to fix non finite pixels PR
- [nodes] Meshing: simplify input depth map folders PR
- [nodes] PanoramaCompositing: add a new graphcut option to improve seams PR
- [nodes] PanoramaCompositing: option to select the percentage of upscaled pixels PR
- [nodes] PanoramaInit: add debug circle detection option PR
- [nodes] PanoramaInit: New parameter to set an extra image rotation to each camera declared the input xml PR
- [nodes] SfmTransfer: New option to transfer intrinsics parameters PR
- [nodes] StructureFromMotion: Add features’s scale as an option PR PR
- [nodes] Texturing: add options for retopoMesh & reorganise options PR
- [nodes] Texturing: put downscale to 2 by default PR
- [sfm] Add option to include 'unknown' feature types in ConvertSfMFormat, needed to be used on dense point cloud from the Meshing node PR
- [ui] Automatically update layout when needed PR
- [ui] Avoid crash in 3D with large panoramas PR
- [ui] Fix graph axes naming for ram statistics PR
- [ui] NodeEditor: minor improvements with single tab group and status table PR
- [ui] Viewer3D: Display equirectangular images as environment maps PR
- [windows] Fix open recent broken on windows and remove unnecessary warnings PR
- [build] Fix cxFreeze version for Python 2.7 compatibility PR
- [ci] Add github Actions PR
- [ci] AppVeyor: Update build environment and save artifacts PR
- [ci] Travis: Update environment, remove Python 2.7 & add 3.8 PR
- [docker] Clean Dockerfiles PR
- [docker] Move to PySide2 / Qt 5.14.1
- [docker] Fix some packaging issues of the release 2019.2.0 PR
- [github] Add exemptLabels PR
- [github] Add issue templates PR
- [github] Add template for questions / help only PR
- [github] Added automatic stale detection and closing for issues PR
- [python] Import ABC from collections.abc PR
For more details see all PR merged: https://github.com/alicevision/meshroom/milestone/10
See AliceVision 2.3.0 Release Notes for more details about algorithmic changes.
Based on AliceVision 2.2.0.
Release Notes Summary:
- Visualisation: New visualization module of the features extraction. PR, New QtAliceVision
- Support for RAW image files.
- Texturing: Largely improve the Texturing quality.
- Texturing: Speed improvements.
- Texturing: Add support for UDIM.
- Meshing: Export the dense point cloud in Alembic.
- Meshing: New option to export the full raw dense point cloud (with all 3D points candidates before cut and filtering).
- Meshing: Adds an option to export color data per vertex and MeshFiltering correctly preserves colors.
Full Release Notes:
- Move to PySide2 / Qt 5.13
- SfMDataIO: Change root nodes (XForms instead of untyped objects) of Alembic SfMData for better interoperability with other 3D graphics applications (in particular Blender and Houdini).
- Improve performance of log display and node status update. PR PR
- Viewer3D: Add support for vertex-colored meshes. PR
- New pipeline input for meshroom_photogrammetry command line and minor fixes to the input arguments. PR PR
- New arguments to meshroom. PR
- HDR: New HDR module for the fusion of multiple LDR images.
- PrepareDenseScene: Add experimental option to correct Exposure Values (EV) of input images to uniformize dataset exposures.
- FeatureExtraction: Include CCTag in the release binaries both on Linux and Windows.
- ConvertSfMFormat: Enable to use simple regular expressions in the image white list of the ConvertSfMFormat. This enables to filter out cameras based on their filename.
For more details see all PR merged: https://github.com/alicevision/meshroom/milestone/9 See AliceVision 2.2.0 Release Notes for more details about algorithmic changes.
Based on AliceVision 2.1.0.
Release Notes Summary:
- 3D Viewer: Load and compare multiple assets with cache mechanism and improved navigation
- Display camera intrinsic information extracted from metadata analysis
- Easier access to a more complete sensor database with a more reliable camera model matching algorithm.
- Attribute Editor: Hide advanced/experimental parameters by default to improve readability and simplify access to the most useful, high-level settings. Advanced users can still enable them to have full access to internal thresholds.
- Graph Editor: Improved set of contextual tools with
duplicate
/remove
/delete data
actions withFrom Here
option. - Nodes: Homogenization of inputs / outputs parameters
- Meshing: Better, faster and configurable estimation of the space to reconstruct based on the sparse point cloud (new option
estimateSpaceFromSfM
). Favors high-density areas and helps removing badly defined ones. - Draft Meshing (no CUDA required): the result of the sparse reconstruction can now be directly meshed to get a 3D model preview without computing the depth maps.
- MeshFiltering: Now keeps all reconstructed parts by default.
- StructureFromMotion: Add support for rig of cameras
- Support for reconstruction with projected light patterns and texturing with another set of images
Full Release Notes:
- Viewer3D: New Trackball camera manipulator for improved navigation in the scene
- Viewer3D: New library system to load multiple 3D objects of the same type simultaneously, simplifying results comparisons
- Viewer3D: Add media loading overlay with BusyIndicator
- Viewer3D: Points and cameras size are now configurable via dedicated sliders.
- CameraInit: Add option to lock specific cameras intrinsics (if you have high-quality internal calibration information)
- StructureFromMotion: Triangulate points if the input scene contains valid camera poses and intrinsics without landmarks
- PrepareDenseScene: New
imagesFolders
option to override input images. This enables to use images with light patterns projected for SfM and MVS parts and do the Texturing with another set of images. - NodeLog: Cross-platform monospace display
- Remove
CameraConnection
andExportUndistortedImages
nodes - Multi-machine parallelization of
PrepareDenseScene
- Meshing: Add option
estimateSpaceFromSfM
and observation angles check to better estimate the bounding box of the reconstruction and avoid useless reconstruction of the environment - Console: Filter non silenced, inoffensive warnings from QML + log Qt messages via Python logging
- Command line (meshroom_photogrammetry): Add --pipeline parameter to use a pre-configured pipeline graph
- Command line (meshroom_photogrammetry): Add possibility to provide pre-calibrated intrinsics.
- Command line (meshroom_compute): Provide
meshroom_compute
executable in packaged release. - Image Gallery: Display Camera Intrinsics initialization status with detailed explanation, edit Sensor Database dialog, advanced menu to display view UIDs
- StructureFromMotion: Expose advanced estimator parameters
- FeatureMatching: Expose advanced estimator parameters
- DepthMap: New option
exportIntermediateResults
disabled by default, so less data storage by default than before. - DepthMap: Use multiple GPUs by default if available and add
nbGPUs
param to limit it - Meshing: Add option
addLandmarksToTheDensePointCloud
- SfMTransform: New option to align on one specific camera
- Graph Editor: Consistent read-only mode when computing, that can be unlocked in advanced settings
- Graph Editor: Improved Node Menu: "duplicate"/"remove"/"delete data" with "From Here" accessible on the same entry via an additional button
- Graph Editor: Confirmation popup before deleting node data
- Graph Editor: Add "Clear Pending Status" action at Graph level
- Graph Editor: Solo media in 3D viewer with Ctrl + double click on node/attribute
- Param Editor: Fix several bugs related to attributes edition
- Scene Compatibility: Improves detection of deeper compatibility issues, by adding an additional recursive (taking List/GroupAttributes children into account) exact description matching test when de-serializing a Node.
See AliceVision 2.1.0 Release Notes for more details about algorithmic changes.
First release of Meshroom.
Based on AliceVision 2.0.0.