.. DO NOT EDIT. .. THIS FILE WAS AUTOMATICALLY GENERATED BY SPHINX-GALLERY. .. TO MAKE CHANGES, EDIT THE SOURCE PYTHON FILE: .. "examples/05_workflows/as_designed_vs_scan.py" .. LINE NUMBERS ARE GIVEN BELOW. .. only:: html .. note:: :class: sphx-glr-download-link-note :ref:`Go to the end ` to download the full example code. .. rst-class:: sphx-glr-example-title .. _sphx_glr_examples_05_workflows_as_designed_vs_scan.py: As-designed vs as-scanned deviation =================================== A real reverse-engineering question: a vendor ships a nominal CAD model of a part, and a depth camera reconstructs the part that actually arrived. Where, and by how much, does the as-built surface deviate from design? This example uses two real models of the same physical object from the T-LESS dataset (Hodaň et al., WACV 2017, CC-BY 4.0): a manually authored CAD model and a model reconstructed from Primesense RGB-D scans. Both are in millimetres in a shared frame. We rigidly align the scan to the CAD nominal, then color the scan by signed distance to the design surface, the first-cut incoming-quality metric. .. GENERATED FROM PYTHON SOURCE LINES 19-27 .. code-block:: Python import numpy as np import pandas as pd import pyvista as pv import pyvista_cad # noqa: F401 # registers the .cad accessor from pyvista_cad.examples import downloads .. GENERATED FROM PYTHON SOURCE LINES 28-29 Load the as-designed CAD model and the as-scanned reconstruction. .. GENERATED FROM PYTHON SOURCE LINES 29-34 .. code-block:: Python cad_path, scan_path = downloads.scan_pair_paths() design = pv.read(cad_path).extract_surface() design .. rst-class:: sphx-glr-script-out .. code-block:: none /home/runner/work/pyvista-cad/pyvista-cad/examples/05_workflows/as_designed_vs_scan.py:31: PyVistaFutureWarning: The default value of `algorithm` for the filter `PolyData.extract_surface` will change in the future. It currently defaults to `'dataset_surface'`, but will change to `None`. Explicitly set the `algorithm` keyword to silence this warning. design = pv.read(cad_path).extract_surface() .. raw:: html

PolyData (0x7f6fa5f36500)
      N Cells:    25614
      N Points:   12806
      N Strips:   0
      X Bounds:   -4.750e+01, 4.750e+01
      Y Bounds:   -2.675e+01, 2.675e+01
      Z Bounds:   -2.950e+01, 2.950e+01
      N Arrays:   3

.. GENERATED FROM PYTHON SOURCE LINES 35-37 The scan carries vertex RGB and normals from the RGB-D reconstruction: real sensor data, not a perturbed copy. .. GENERATED FROM PYTHON SOURCE LINES 37-41 .. code-block:: Python scan = pv.read(scan_path) scan .. raw:: html

PolyData (0x7f6fa5f341c0)
      N Cells:    42901
      N Points:   21852
      N Strips:   0
      X Bounds:   -4.838e+01, 4.815e+01
      Y Bounds:   -2.843e+01, 2.761e+01
      Z Bounds:   -3.065e+01, 2.919e+01
      N Arrays:   2

.. GENERATED FROM PYTHON SOURCE LINES 42-45 Rigidly register the scan onto the CAD nominal. Real scans land in an approximate pose; the deviation is only meaningful after an ICP alignment removes the residual rigid offset. .. GENERATED FROM PYTHON SOURCE LINES 45-49 .. code-block:: Python scan = scan.align(design) scan .. raw:: html

PolyData (0x7f6fa5f340a0)
      N Cells:    42901
      N Points:   21852
      N Strips:   0
      X Bounds:   -4.838e+01, 4.821e+01
      Y Bounds:   -2.823e+01, 2.808e+01
      Z Bounds:   -3.061e+01, 2.983e+01
      N Arrays:   2

.. GENERATED FROM PYTHON SOURCE LINES 50-52 Compute deviation: signed distance from each scan point to the CAD surface. .. GENERATED FROM PYTHON SOURCE LINES 52-68 .. code-block:: Python scan = scan.compute_implicit_distance(design) dev = scan['implicit_distance'] pd.DataFrame( { 'metric': ['min', 'max', 'mean', 'RMS'], 'deviation_mm': [ dev.min(), dev.max(), dev.mean(), np.sqrt((dev**2).mean()), ], } ) .. raw:: html

	metric	deviation_mm
0	min	-11.393228
1	max	4.331330
2	mean	-0.385970
3	RMS	2.173232

.. GENERATED FROM PYTHON SOURCE LINES 69-73 Side by side: the as-designed CAD on the left, the registered scan colored by signed deviation on the right. Red is proud of nominal, blue is below it; the bulk of the surface is within scanner noise, with clear excursions where the reconstruction is incomplete. .. GENERATED FROM PYTHON SOURCE LINES 73-89 .. code-block:: Python clim = float(np.percentile(np.abs(dev), 95)) pl = pv.Plotter(shape=(1, 2)) pl.subplot(0, 0) pl.cad.add(design, color='lightgray') pl.subplot(0, 1) pl.add_mesh( scan, scalars='implicit_distance', cmap='coolwarm', clim=(-clim, clim), scalar_bar_args={'title': 'deviation [mm]'}, ) pl.link_views() pl.show() .. tab-set:: .. tab-item:: Static Scene .. image-sg:: /examples/05_workflows/images/sphx_glr_as_designed_vs_scan_001.png :alt: as designed vs scan :srcset: /examples/05_workflows/images/sphx_glr_as_designed_vs_scan_001.png :class: sphx-glr-single-img .. tab-item:: Interactive Scene .. offlineviewer:: /home/runner/work/pyvista-cad/pyvista-cad/doc/examples/05_workflows/images/sphx_glr_as_designed_vs_scan_001.vtksz .. rst-class:: sphx-glr-timing **Total running time of the script:** (0 minutes 0.969 seconds) .. _sphx_glr_download_examples_05_workflows_as_designed_vs_scan.py: .. only:: html .. container:: sphx-glr-footer sphx-glr-footer-example .. container:: sphx-glr-download sphx-glr-download-jupyter :download:`Download Jupyter notebook: as_designed_vs_scan.ipynb ` .. container:: sphx-glr-download sphx-glr-download-python :download:`Download Python source code: as_designed_vs_scan.py ` .. container:: sphx-glr-download sphx-glr-download-zip :download:`Download zipped: as_designed_vs_scan.zip `