Request: Dense Cloud Filtering

[Marcel]

We are struggling with building the mesh for projects with very big Dense Clouds (200 million points or more).

The problem is that the Build Mesh stage always builds the mesh using the entire point cloud. So even if we build a mesh with a 80 million polygon target, it will first build the mesh with the maximum amount of polygons (using all points in the cloud), and only decimate afterwards. This takes a lot of time and we often run out of memory, even with 128GB.

What would be a very good solution, is if we could optimize/decimate the Dense Cloud. For example, in MeshLab you can use "Poisson Disk Sampling", this function filters the point cloud and removes (unnecessary) points.

We would love to have such a function in Photoscan. This would save us a lot of time, and depending on the filtering algorithm it could result in a mesh that is the same quality as the unfiltered Dense Cloud.

[andyroo]

+1 for dense point cloud filtering.

I could see different filters being useful for different applications.  It would be nice to have several options for filtering.

For example, I have been using LASTools and Point Cloud Library to either decimate (random or every nth point) or filter (statistical outlier removal, radius filter) the dense pointclouds before DSM production. But being able to do these operations before generating the mesh would be great!

Sometimes I have dreams where I am using a filter that allows me to write equations based on variables in the pointcloud, on individual points or within a radius. In those dreams I make beautiful rasters and solve all of my problems by doing things like conditional filtering based on RGB values, and filtering by intensity from the gamma-corrected RGB luminance. And filtering by precision values derived from the triangulation uncertainty..

wow. that would be amazing.

[dcm39]

+1 more for dense cloud filtering/decimating.

At present, if I want to use the point cloud I filter it with the GMT blockmean command. But it would be really useful to be able to do this before the meshing in the photoscan workflow.

Other programs also aren't really able to maintain the vertex colour  etc... again, would be useful!

[Marcel]

Thanks for your +1.

I've been thinking about this the last few weeks, and the reason why I feel Dense Cloud Filtering is important is because currently we have no way to predict how many points the Dense Cloud phase creates. It takes 20+ hours to process, and only afterward you see if it has the right number of points. If it does not have enough points then we don't have the detail we need, if it has too many points then meshing it becomes really hard.

Some of our recent projects have Dense Clouds between 1 billion and 2 billion points. The amount of detail is amazing, but we really struggle to create meshes from these Clouds. We can split them up in several chunks, but this adds a lot of extra work.

Being able to reduce the number of points in the Dense Cloud would make it so much easier. If the reduction is done in a smart way (with only redundant points being removed) then this would improve quality  because we could always build the Dense Cloud at the maximum level without having to be afraid of un-meshable Clouds.

[Porly]

+1 for dense cloud filtering and smoothing option for marked faces

[wojtek]

+1 from me as well.

Are there any plans/developments in regards to this feature?

[taran]

+1 from me - deeply useful for those working with huge datasets.

[Bruno Andrieu]

+2 for me !!

[gregoconn]

+1 for me. marcel explains the reasoning perfectly.

I've been thinking about this the last few weeks, and the reason why I feel Dense Cloud Filtering is important is because currently we have no way to predict how many points the Dense Cloud phase creates. It takes 20+ hours to process, and only afterward you see if it has the right number of points. If it does not have enough points then we don't have the detail we need, if it has too many points then meshing it becomes really hard.

Being able to reduce the number of points in the Dense Cloud would make it so much easier. If the reduction is done in a smart way (with only redundant points being removed) then this would improve quality  because we could always build the Dense Cloud at the maximum level without having to be afraid of un-meshable Clouds.

[MartinBim]

+1 as marcel said

[3DFranz]

+1

[Mohammed]

+++++1 great idea and i'll save our time alot.

[MHThomsen]

+1 from me too. Much needed feature.

[YesOk]

+1

No need to save point cloud color data if it is easier to filter.

[GeoGecco]

+1 for me