Point Cloud to Mesh
Point cloud triangulation turns a set of 3D points captured by LiDAR or other scanning technologies into a continuous triangle mesh.
Discover morePoint Cloud Processing Library for C++ and Python
From raw scans to a usable 3D model
MeshLib takes point clouds from LiDAR and 3D scanners through every step to a finished model: import, clean, sample, estimate normals, align scans with ICP and reconstruct a triangle mesh. One SDK for C++, Python, C# and C.
Point cloud algorithms in MeshLib
Mesh Triangulation
Mesh triangulation converts point clouds into triangle meshes quickly, using vertex normals to guide the surface.
Discover more
ICP: Iterative Closest Point
ICP aligns two 3D datasets: it takes a source point cloud or mesh and estimates the rigid transformation that places it onto the target.
Discover moreHow does MeshLib process a point cloud?
A typical scan-to-model pipeline has six steps. Each one is a MeshLib function you can call on its own or chain with the others.
Import scans
Load point clouds from E57, LAS, LAZ, PLY, PTS, XYZ, XYZN, ASC, CSV, OBJ and DXF files.
PointsLoad::fromAnySupportedFormatRemove outliers and sample
Find stray points far from the surface, then thin dense scans with uniform or grid sampling to keep the detail you need at a size you can process.
findOutlierspointUniformSamplingpointGridSamplingEstimate and orient normals
Raw scans usually come without normals. MeshLib computes them and orients them consistently, which reconstruction and point-to-plane registration depend on.
makeOrientedNormalsorientNormalsAlign scans with ICP
Register overlapping scans with point-to-point or point-to-plane ICP, or align many scans at once with multiway ICP. ICP is a local refinement: it starts from a rough initial placement of the scans.
ICPMultiwayICPReconstruct a mesh
Triangulate the points directly, or fuse them through a voxel grid to get a single surface from overlapping scans.
triangulatePointCloudpointsToMeshFusionMeasure and compare
Compute distances between point clouds and meshes to check a scan against a reference model.
findMaxDistanceSqfindSignedDistances
Which point cloud formats does MeshLib read?
Laser scanner and LiDAR formats load directly, without a conversion step.
- E57
- LAS
- LAZ
- PLY
- PTS
- XYZ
- XYZN
- ASC
- CSV
- OBJ
- DXF
Point cloud to mesh in Python and C++
Three calls take a scan from a file to a triangle mesh. The same API is available in C# and C.
from meshlib import mrmeshpy as mm
cloud = mm.loadPoints("scan.ply")
mesh = mm.triangulatePointCloud(cloud)
mm.saveMesh(mesh, "scan_mesh.ply")// error handling omitted for brevity
auto cloud = MR::PointsLoad::fromAnySupportedFormat( "scan.ply" );
auto mesh = MR::triangulatePointCloud( *cloud );
MR::MeshSave::toAnySupportedFormat( *mesh, "scan_mesh.ply" );Try MeshLib now
Booleans, mesh repair, decimation and ICP — in your browser, as JavaScript and WebAssembly. No install, no sign-up.
Open live demoOr install the SDK — free to evaluate
Point cloud processing FAQ
What is point cloud processing?
Point cloud processing is the set of operations that turn raw 3D scan points into usable data: loading the scan, removing outliers, sampling, estimating normals, aligning several scans and reconstructing a mesh surface.
Can MeshLib convert a point cloud to a mesh?
Yes. MeshLib triangulates point clouds directly with triangulatePointCloud and can also fuse overlapping scans through a voxel grid with pointsToMeshFusion. See Point Cloud to Mesh.
Which point cloud formats does MeshLib support?
MeshLib reads E57, LAS, LAZ, PLY, PTS, XYZ, XYZN, ASC, CSV, OBJ and DXF point clouds. The full list is on the file formats page.
Does MeshLib register (align) 3D scans?
Yes. MeshLib aligns scans with point-to-point and point-to-plane ICP and aligns many scans at once with multiway ICP. Point-to-plane ICP needs normals, so compute them first if the scan has none. See ICP.
Can I use MeshLib for point clouds in Python?
Yes. The Python package exposes the same point cloud functions as the C++ core; install it with pip install meshlib.