Skip to content

Structured Point Clouds

In one sentence: Triangulate a structured (grid) point cloud directly into a mesh by connecting neighbouring grid cells into triangles.

Overview

Workbench: Reverse Engineering
Menu: Reverse Engineering → Surface Reconstruction → Structured Point Clouds
Shortcut: none

Converts a Points::Structured object (a Width × Height ordered grid of points) into a triangle mesh by connecting adjacent grid cells. No surface fitting is performed — the mesh simply links the grid points with edges. Invalid cells (NaN coordinates) result in holes in the mesh.

Intuition

A structured point cloud is like a digital photograph: each pixel has a row and a column. Structured Point Clouds connects each pixel to its four neighbours to form two triangles per grid cell, giving you a mesh that is pixel-accurate — every input point becomes a vertex, and the mesh topology exactly mirrors the sensor's scan grid.

When to use it

  • Your scan comes from a structured sensor (depth camera, structured-light scanner, spinning LIDAR) and is already loaded as a Points::Structured.
  • You want the fastest possible mesh that exactly preserves all scan points.
  • You are doing depth-image processing where the grid topology matters.

When NOT to use it

  • Unstructured point clouds — this tool requires Points::Structured input. For unstructured clouds, use Poisson Reconstruction instead.
  • Noisy or sparse sensor data — directly triangulating a noisy grid produces a very rough, bumpy mesh. Consider smoothing the mesh afterwards with Mesh workbench tools.

Step-by-step

  1. Ensure you have a Points::Structured object (created via Points → Structured Point Cloud, or imported from a structured format).
  2. Select the Points::Structured object.
  3. Choose Reverse Engineering → Surface Reconstruction → Structured Point Clouds.
  4. FreeCAD creates a Mesh::Feature named "View mesh".

Parameters

This tool has no parameters — the triangulation is fully determined by the Width × Height grid structure of the input Points::Structured.

NaN handling

Grid cells where the sensor produced no valid return are stored as (NaN, NaN, NaN) in the structured cloud. Any triangle that would include a NaN vertex is skipped, leaving a hole in the mesh at that grid location.

Common mistakes and pitfalls

Tool fails: object is not Points::Structured

Cause: You selected a plain Points::Feature (unstructured cloud) instead of a Points::Structured.
Fix: First convert the cloud to structured via Points → Structured Point Cloud, then apply this tool.

Resulting mesh is extremely bumpy

Cause: The raw scan has sensor noise at every grid cell, which the direct triangulation faithfully reproduces.
Fix: Smooth the mesh after triangulation using Mesh workbench → Smoothing.

Python API

import ReverseEngineering as Reen
import FreeCAD as App

doc = App.ActiveDocument
structured = doc.getObject("StructuredCloud")  # Points::Structured

mesh_feat = doc.addObject("Mesh::Feature", "ViewMesh")
mesh_feat.Mesh = Reen.viewTriangulation(
    Points=structured.Points,
    Width=structured.Width,
    Height=structured.Height
)
doc.recompute()

See also