Open source software, remote sensing, computer vision, machine learning, #pointcloudpipe
Joined July 2008
- Tweets121
- Following171
- Followers71
- Likes107
Come and see my talk at 16:45 today to learn about semantic segmentation of 3D point cloud data! talks.osgeo.org/foss4g-2022/… #FOSS4G2022 @pointcloudpipe
Bradley J Chambers retweeted
Newly announced for #FOSS4G2022... viewer.copc.io now supports drag-n-drop of local copc.io files! Thanks @TheMrCam and @connor2n2o from the @hobuinc team for the wizardry. Convert your data to COPC with @pointcloudpipe or @qgis and share away!
Bradley J Chambers retweeted
Thanks @tiledb for contributing PDAL's enhanced Python API. Thanks to this, you can now compose PDAL operations in Python instead of just passing JSON to PDAL. It should open up huge possibilities for people mixing point clouds with Python's ML machinery.
Thanks to @tiledb, @pointcloudpipe now has an enhanced Python API that support composition, meshio, streaming, and better numpy support. Grab the source from pypi.org/project/PDAL/ or wait for @condaforge to catch up its feedstock.
1/ It can be difficult to effectively visualize large numbers of classes in point cloud data, especially in the presence of class imbalance.
4/ All classes not currently receiving focus are displayed using a mixture of intensity and ambient occlusion. In doing so, we are able to maintain the context provided by the full scene while evaluating individual classes (cars).
5/ "Autzen" point cloud courtesy @pointcloudpipe used under CC-BY-4.0 / enhanced using custom PDAL Python scripts and @CloudCompareGPL.
1/ Lidar intensity values often add meaningful context, but sometimes these values, at least for visualization purposes, are not ideal.
4/ Finally, the product of the histogram equalization and ambient occlusion values can be computed to arrive at a visually appealing rendering of the data.
5/ "Autzen" point cloud courtesy @pointcloudpipe used under CC-BY-4.0 / enhanced using custom PDAL Python scripts and @CloudCompareGPL.
Experimenting with different ways to visualize categorical labels in point cloud data. Autzen LAZ courtesy @pointcloudpipe used under CC-BY-4.0 / 12-class Paired color scheme enhanced with ambient occlusion using @CloudCompareGPL. grovercsllc.com/
Bradley J Chambers retweeted
Thanks to Paul Harwood from runette.co.uk/ and Jaime Soto from @luminartech, PDAL now has a C API released and available on @condaforge anaconda.org/conda-forge/pda…
Bradley J Chambers retweeted
Less than a week left on our crowdfunding campaign to support point cloud data in #QGIS. Note that contributions are still welcome and will be used towards stretched goals, see:
lutraconsulting.co.uk//crowd…
Bradley J Chambers retweeted
Crowdfunding target reached for point clouds in #QGIS - huge thanks for your support so far. We are preparing stretch goals, stay tuned! lutraconsulting.co.uk/crowdf…
Bradley J Chambers retweeted
Incredible response from the community with 4/5th of the funds already pledged. Just 10k more and we can start the implementation. Who wants to be on the list of donors - there are still a few places left: lutraconsulting.co.uk/crowdf…
Bradley J Chambers retweeted
PDAL 2.2.0 is now released. github.com/PDAL/PDAL/release… New features include 'where' syntax for directed processing, OBJ reader support, @Esri I3S/SLPK performance improvements, and @entwineio reader supports accelerated polygon-bounded queries.
