koide3

I added instructions about partial rendering that allows reducing the rendering cost of large static point clouds by accumulating partially rendered points over time. Note that it may not work...

Thanks for reporting the issue. I'll fix the first one by checking both `glm_FOUND` and `GLM_FOUND`. I'll update `CMakeLists` to resolve both soon.

You might forget enabling dynamic points integrator https://koide3.github.io/direct_visual_lidar_calibration/example/#2-preprocessing_1.

It's my bad. Here `params.voxel_resolution` is the voxel resolution for final downsampling (e.g., 0.05) while the first argument for `iVox` is the voxel resolution for data association (e.g., 1.0).

Thanks for your request. The current implementation of downsampling considers only point coordinates and ignores other attributes. I think we can add a slight modification to support point colors. Maybe...

MacOS support sounds so nice. Let me know when the PR gets ready to be reviewed.

`deskewed` has only covariance matrices and lacks normals. You need to add normals by calling `add_normals` if you want to use CT_ICP. There is also CT_GICP that uses covariances matrices,...

You can let `covariance_estimation` estimate normals in addition to covs: ```cpp std::vector normals; std::vector covs; covariance_estimation.estimate(deskewed->points_storage, neighbors, normals, covs); auto deskewed = std::make_shared(factor->deskewed_source_points(values)); deskewed->add_normals(normals); deskewed->add_covs(covs); deskewed->add_intensities(points->intensities, points->size()); target_ivox->insert(*deskewed); ``` It...

Thanks, I'll check if it works on older versions. If not, I'll add some fallback.

The image doesn't need to be rectified, this package does it internally. The camera info data is saved in a json file, and programs after preprocessing use the saved camera...