CHANGELOG.md

Changelog

0.10.2 New console script cars-devibrate (December 2024)

Added

• Add cars-devibrate script (based on the work of Julien Michel)

Fixed

• Fix classification holes in epipolar geometry [#949]

0.10.1 Using importlib instead of setuptools_scm (December 2024)

Fixed

• Fix wheel by using importlib instead of setuptools_scm
• Cast color before fusion to avoid asymetrical behavior [#954]
• Adapt cars-starter to the new CARS interface
• Documentation corrections

0.10.0 Unique pipeline (December 2024)

Added

• Add bulldozer-dtm dependency to cars with a new bulldozer base DSM filling application [#901]
• Add cars-filter dependency with outlier filtering functionalities in epipolar geometry [#867]
• New parameter use_cross_validation in dense matching to manage Pandora cross validation [#900]
• New parameters altitude_delta_min and altitude_delta_max in initial_elevation [#902]

Changed

• Cars pipelines have been replaced by a "default" pipeline [#895]
• Support for Python version superior than 3.10 with libsgm 0.5.0 [#912]
• Support for Python 3.8 is over [#930]
• The no merging mode is now the operational mode in cars, outlier filtering is done in epipolar geometry and laz are now saved in triangulation application [#895]
• Add an index.json file in cars official product [#893]
• Improve formatting of cars output products, improve cleaning of the output directory [#874]
• Improve SIFT default configuration [#928]
• Optimize SIFT filtering and grid correction [#869]
• Optimize classification resampling [#835]
• Simplify the check input function [#920]
• Support z_inf and z_sup in depth map inputs [#897]
• User and developer documentation updates [#935] and [#885]
• Use shareloc 0.2.4

Fixed

• Typo fixes (with API changes) [#934]
  • dense_matches_filling application renamed into dense_match_filling
  • holes_detection application renamed into hole_detection
  • point_cloud_outliers_removing renamed into point_cloud_outlier_removal
  • points_cloud renamed into point_cloud in the code
• Fix crashes during report generation [#916]
• Fix cars when using image with negative y axis [#924]
• Fix no-data handling in elevation data [#922]
• Make cars robust to different classification geotransforms [#923]

0.9.0 Output interfaces (September 2024)

Added

• Usage of up/down sampled sensor images [#797]
• Pleiades usage tutorial [#792]

Changed

• Refactoring of input and output configuration [#824]
• Choose interpolation mode used in resampling [#831]
• Displacement grids changed to localisation grids [#828]
• Cars-resample and Cars-rasterize optimisation [#803]
• Dsm radius parameter can now be used as a float [#792]
• Refactoring of geoid usage [#830]

Fixed

• Artefacts in performance map [#846]
• Performance map modified [#898]
• DSM clipping removed in notebook [#853]
• Dask / pandas future warning removed [#759]
• Crash avoided when temporary files lost [#847]
• Blue pixels in orthorectified image [#812]

0.8.0 Orfeo Toolbox dependency removal (June 2024)

Added

• Auto mode for orchestrator adapts to machine resources [#798]

Changed

• Orfeo Toolbox geometry plugin is no longer available [#805]
• Exogenous DEM is cropped before disparity grid generation to reduce time and memory consumption [#827]

Fixed

• Subpixellic shift during resampling is fixed [#799]
• Input mask can be saved as output of dense_point_clouds_to_dense_dsm pipeline [#816]
• Ambiguity is normalized globally and no longer by tiles [#810]
• Classified pixels outside of epipolar footprint are no longer filled [#814]
• Statistics on confidence intervals are added and warnings are removed [#817]
• Parameter check_inputs can be used with Shareloc geometry plugin [#820]
• Masked and filled pixels are no longer masked in final DSM [#836]
• Cache of GDAL is limited to 500Mb to avoid memory overload during resampling [#823]

0.7.6 Confidence intervals (April 2024)

Added

• Integration of Pandora confidence intervals [#764]
• Option to choose endogenous or exogenous DEM to resample images [#790]
• Geomodel parameter is optional [#785]
• Path to JSON file is allowed for dense matching loader configuration [#552]
• Support for point cloud denoising plugin [#771]
• Check footprint of given initial elevation [#780]

Changed

• Version 0.2.0 of Shareloc is used : fast grid generation [#781]

Fixed

• Invert rows and columns for generation of disparity range grids with Shareloc [#804]
• Better typing of output files [#758]
• Faster LAZ point cloud saving [#795]
• Avoid computation of useless epipolar tiles when a ROI is given for no_merging pipelines [#791]
• Compute footprint of endogenous DEM according to images and epipolar footprint [#780]
• Broadcast delayed objects with dask to avoid memory overload [#796]

0.7.5 Pandora integration (March 2024)

Added

• New default pipeline sensors_to_dense_dsm_no_merging with cumulative rasterisation [#698]
• Profiling report for multiprocessing mode [#745]
• Dashboard showing tiles processing during a run [#765]
• Save of sparse matches in sensor geometry [#761]
• Relaunch of frozen tasks after timeout [#768]

Changed

• Version 1.6.0 of Pandora is used [#688]
• Version 0.2.0a2 of Shareloc is used [#756]
• Sparse matching processed by strip to improve computation time [#753]
• DEM generation improved : less artifacts, filling of invalid regions [#754]
• Default orchestrator is multiprocessing [#755]
• Disparity interval cropped when it is too high [#757]

Fixed

• Data is typed to save memory consumption [#661]
• Preload of CARS in workers for multiprocessing mode to reduce forking time [#730]
• CPU usage limited to 100% per worker [#750]
• Output classification pixels cannot be floating-point numbers anymore
• ROI not used for endogeneous DEM [#772]

0.7.4 Local disparity ranges (December 2023)

Added

• Local disparity ranges are used [#695]
• Write point cloud by pair [#727]
• Left and right classification fusion on disparity map [#729]

Changed

• Plane filling method: no propagation by default [#725]

Fixed

• Slurm cluster launch [#736]
• None tiles managed in save_cars_dataset [#716]
• Color data type propagated to the end [#642]
• Memory unreleased in multiprocessing mode [#731]
• Loglevel could not be set [#734]

0.7.3 CARS fully pip installable (November 2023)

Added

• CARS can now be pip installed [#639]
• Name of color bands propagated in point cloud [#696]

Changed

• Slurm account now as a parameter of Orchestrator [#712]
• pkg_resources not used anymore [#700]

Fixed

• Rasterization in epsg:4326 supported [#724]
• Estimation of DSM roi in Shareloc plugin [#701]

0.7.2 Self calculated initial elevation (October 2023)

Added

• Multiprocessing cluster: add factorization of Delayed [#673]
• 8 bits images are managed with particular sift parameters [#668]
• Self calculated initial elevation with sift matches [#664]
• Add debug mode with roi using initial tiling [#475]
• Information map about dense matches filling [#679]
• Script to generate CARS configuration [#678]

Changed

• Multi points direct location [#708]
• Application parameter "method" now optional [#692]
• Geometry plugin interface [#623]
• Nodata parameter now optional [#296]

Fixed

• Orchestrator used does not match the one reported in the logs [#681]
• used_conf is no longer created when the application is initialized [#693]
• Parameter "mask" can be used in "dense_point_clouds_to_dense_dsm" [#722]

0.7.1 Slurm Cluster (July 2023)

Added

• Dask Slurm cluster [#624]
• Crop disparity range with thresholds [#626]
• Used source information map generated [#648]

Changed

• Write workers logs in one single file [#659]
• Use tempfile to generate default orchestrator folder instead of /tmp [#663]

Fixed

• Cython upgrade bug [#674]
• Too much used disk memory in MP mode [#632]

0.7.0 CARS installable without OTB (June 2023)

Added

• Notebook: add masks and classification options [#615]
• Option: Set python interpreter to use in PBS Dask Cluster [#611]
• DensePointCloudToDenseDSM Pipeline: now re-entrance with confidence and classification [#602]
• Shareloc Geometry plugin is now internal [#618]
• Use_sec_disp option allowing to use right disparity map is removed [#638]
• CARS can now be installed without OTB [#637]
• Epipolar tiles corresponding to region outside sensor images are removed before any computation [#635]
• Add CARS progress messages [#658]

Changed

• Multiprocessing cluster doesn't freeze when bug occur anymore [#607]
• Resampling is not done with OTB anymore [#508]
• Pyproj is no longer constrained [#646]

Fixed

• Notebook: retrieve dsm [#614]
• Doc: Roi parameter disappearing [#616]

0.6.0 A new input interface (mask and classification) and new pipelines (March 2023)

Added

• All confidence maps from Pandora are propagated [#521]
• New pipelines: sensor to point cloud, point cloud to dsm[#567]
• Generate performance map [#590]
• Density parameter for terrain tile size computation in pc to dsm pipeline [#566]
• Multiprocessing mode available for DenseMatchesFilling [#539]
• Rasterization coded in C++ [#523]

Changed

• P+XS Fusion is no longer available in CARS [#579]
• Input mask is now a binary validity mask used in the whole pipeline [#566]
• Region of Interest with GeoJson or Shapefile [#580, #242]
• Refactoring of Masks, Classifications and Color [#577, #578]
• Move the former "set_to_ref_alt" option to DenseMatchesFilling application [#577, #578]

Fixed

• Bug DenseMatchesFilling Plane [#599]

0.5.4 Grid correction in the configuration file and a better estimate of disparity range (February 2023)

Added

• A better estimate of disparity range [#538]
• Prepare and use Epipolar A Priori data: grid correction and disparity range [#458]
• Use custom DASK configuration for CARS [#554]
• Use adaptative DASK cluster [#541]
• Update Pandora to 1.3 [#459][#575][#587]
• Point cloud saving in the low resolution pipeline [#498]

Changed

• Update black formatting rules [#556]
• Update dask memory logger, disable it by default [#527]
• OTB_MAX_RAM_HINT not used for tile sizing anymore [#541]
• Deactivate DASK dashboard by default [#343]

Fixed

• Line artefact on right disparity map [#525]
• Pipeline configuration override [#533]
• Scipy an xarray broken interfaces [#561]
• Set masked pixels to nodata in point clouds from triangulation [#465]
• Set log level of dask cluster to warning [#545]

0.5.3 Points clouds as LAZ files (December 2022)

Added

• Dump point clouds as LAS [#438]
• Refactoring of tiling grids and margins [#451]
• Generate a runnable full configuration file [#422]
• Add hole filling application for disparity maps [#275]

Changed

• Add minimum number of sift matches in configuration [#514]

Fixed

• Fix pylint errors (no-member) in application template [#437]
• Fix bug PBS Dask [#515]
• Fix wrong images saved in notebooks
• Fix bug in disparity map saving [#484]

0.5.2 Profiling and dense matching confidence information (November 2022)

Added

• Add ambiguity transmission through pipeline [#478]
• Add memory/cpu profiling mode [#473]
• Clean notebooks outputs [#497]

Changed

• Clean setup.cfg with optionnal package data [#467]
• Clean disk data in Multiprocessing mode [#454]

Fixed

• Fix dimension bug during point cloud fusion [#482]

0.5.1 Minor fixes post refactoring (September 2022)

Changed

• New User Guide version [#448]

Fixed

• Fix dask upgrade bug [#436]
• Fix temporarily local path to CARS OTB installed apps [#434]
• Fix multiprocessing bug [#446]
• Fix (remove) warning: tile ignored [#455]

0.5.0 Refacto CARS (August 2022)

Be careful, this version changes API and configuration ! See new documentation.

Added

• Add documentation generation pre-commit when git push [#406]
• Add docker images files for jupyter and cars tutorial [#418]
• Add cars tutorial main_tutorial as notebook slides [#421]
• Add pandora mccn simplified installation process [#426]

Changed

• Refactoring of all CARS code by implementing new concepts as application, CarsDataset, orchestrator, pipeline and configuration. See documentation for more details. [#332]
• upgrade support to python 3.8 only [#372]
• change quick_start and notebooks to tutorials structure [#419]

Fixed

• Remove setuptools pre installation [#341]
• Remove click 8.0.4 dependency [#341]
• Add automatic use of setup.cfg docs extra_require for readthedocs conf [#367]

0.4.2 OTB/GDAL upgrade (July 2022)

Fixed

• Fix gdal-config dependency in docker with rasterio upgrade [#404]
• Upgrade OTB to 7.4.1 in Dockerfile [#404]

0.4.1 Stabilization - Upgrade Pandora 1.2.* (June 2022)

Added

• Add Authors file for contributors [#400]

Changed

• Upgrade to pandora 1.2.* [#371]
• Improve performance of CARS cli with no argument [#357]
• Add egm96 internal geoid hdr header [#335]
• Cluster dask code structure evolution [#355]
• Clean : remove hard coded indexes in the cloud fusion [#394]
• Gitlab template evolution [#399]
• Makefile evolution and clean [#391]

Fixed

• Fix min_elevation_offset option when user_disp_min is greater than disp_min [#348]
• Temporary bugfix by forcing click version to avoid dask.distributed pbs cluster trouble [#383]
• Fix CI errors with pylint upgrades and docker apt package ubuntu upgrades [#390, #388, #385, #384, #380, #379, #376, #369, #368, #339]
• Monitor pytest CI performance with debug information [#373]
• Fix shapely version upgrade bug [#349]
• Fix pygdal setuptools version upgrade bug [#333]
• Fix OTB geometry plugin bugs [#393, #396]
• Fix Loader geometry bug: non static schemas management [#395]
• Filter useless verbose dask warning [#353]

0.4.0 Geometry Loader (September 2021)

Added

• Write used dask configuration in output directory. [#224]
• Add argparse file option @"opts.txt" + doc [#265]
• Add Contributor Licence Agreement [#257]
• Add quality code tools isort, black, flake8 [#247, #269, #271]
• Add prepare/compute_dsm notebook [#246]
• Add sonarqube configuration [#198]
• Add Geometry Loader mechanism [#287]
• Add OTBGeometry loader [#287]

Changed

• Update/Clean package setup (add Makefile, clean requirements) [#210, #193, #305, #197]
• Make pip install -e work [#207]
• Update/Clean environment parameters [#166, #251]
• Move cars_cli.py from bin to cars. [#188, #203]
• Rename cars_cli to cars. [#188]
• Change default nb_workers to 2 [#218]
• Allow multiprocessing fork mode. [#283]
• Force OpenMP use in dask, and TBB in multiprocessing. [#304]
• Change loglevel argument API to pipeline level [#310, #311]
• Upgrade and fix pandora dependency [#235, #267, #274, #273, #309, #188]
• Clean quality code pylint and sonarqube conf [#302, #209]
• Change CARS loglevel default to WARNING + clean output [#239, #300, #143]
• Refactoring CARS file/module organization [#216, #259]
• Clean Documentation + ReadTheDocs [#160, #258]
• Dockerfile update [#219]
• Update/Clean README [#200]
• Geoid file indicated via the static configuration [#287]

Fixed

• Fix epipolar size. [#206, #237, #248]
• Fix and clean tbb support [#267, #276, #304]
• Fix align_with_lowres_dem in mp mode [#286]
• Fix Delaunay algorithm between epi and terrain tiles [#277]
• Fix dask ComputeDSMMemoryLogger api [#202]
• Handle margins when setting the disparity to 0 [#201]
• Fix car cli with setuptools_scm version [#199]

0.3.0 Multi-classes mask management (December 2020)

Added

• Mask management : change input format and internal behaviors. [#147, #170]
• Constants.py added to optimize code [#172]
• Default elevation option [#111]
• Satellite angles information on prepare step [#58, #190]
• Changelog added [#185]

Changed

• Update rasterization tests [#177]
• Integrate a new memory estimation method [#158]
• Version handling with setuptools_scm [#194]

Fixed

• Fix CLI parsing for ROI option : CLI API changes (-i for input json file and -o for output directory) [#93, #174]
• Pylint code clean [#191]
• Add constants.py file [#172]

0.2.0 First Open Source Official Release (July 2020)

• 3D functional pipeline : Python3 and C++ OTB modules
• Python API
• CLI command
• Documentation basics
• Continuous Integration on unit tests