Existing building

11Point cloud to building model

Turn measured points into identifiable building elements useful for the design.

Proposed organisation · the numbered catalogue follows the latest reorganisation plan. Contributions are planned; this view does not confirm operational activation.

One stage, complementary contributions.Back to the decades

Contributing to the stage

Expertise working together.

Each role retains its own objective and deliverables. Open the details to read how it contributes to this stage.

EC11Planned contribution

Point cloud to building model

Turn measured points into identifiable building elements useful for the design.

Method and deliverables

How it contributes

Combines point clouds with photographs and plans to interpret surfaces, rooms and openings. Measured regions, inferred completions and unresolved areas remain distinguishable in the native IFC.

Expected deliverables

  1. Partial or scoped as-is IFC
  2. Source comparisons and uncertainty views

“Registered” describes the role in the catalogue. Deliverable availability and project status are verified separately.

EC11 · THE METHOD IN OUR HOUSES

From points to a readable building.

Follow observations, decisions and checks at Casa del Noce and Casa del Forno: eleven steps explaining why BIM reconstruction takes more than a convincing image.

Start with the survey

Which parts of the building were actually observed?

Observations, decision and outcome

The first reading covers the point set as a whole: overall shape, height distribution, dense areas and areas without returns. The survey retains irregularity, obstacles and occlusion. These differences remain visible even when they make the result less regular than an idealised model.

The decision

We retain the relationship with the original observations and distinguish exploratory views from the measurements used in reconstruction. A gap in the cloud initially means an unobserved area, not a window or a demolition.

The outcome of this step

A common geometric context with identifiable observations and gaps.

Further reading: Tang et al., 2010Gourguechon et al., 2022

Identify the spaces

Which house, room and wall does each observation belong to?

Observations, decision and outcome

Photographs, historical documents and survey views are compared to identify rooms. At Forno, the common room and rooms at different heights make this step especially important. At Noce, the external stair connects the reading of the façade with the upper apartment.

The decision

Documentary plans help identify relationships between rooms; they do not override surveyed geometry. Photographs clarify context without automatically becoming a source of measurements. Codes containing n and f distinguish the houses; the level remains a separate attribute.

The outcome of this step

Stable identities for spaces and elements, even when their representation changes.

Further reading: Tang et al., 2010

Read levels and connections

Where do rooms start and end, particularly across split levels?

Observations, decision and outcome

Survey plans at different heights and longitudinal sections reveal floors, soffits, passages and level changes. At Forno, the room groups cannot be reduced to a single horizontal floor. At Noce, sections also expose lower rooms that still need reconstruction.

The decision

We retain local levels and observed discontinuities. An upper surface and a soffit do not by themselves establish a floor build-up. A connection between two rooms requires a documented passage, not simply proximity.

The outcome of this step

A spatial reading that respects level changes, room sequences and missing parts.

Further reading: Macher et al., 2017Gourguechon et al., 2022

Separate observed faces

Does a visible surface really describe the entire wall?

Observations, decision and outcome

The survey contains wall faces, finishes, returns and nearby objects. For the irregular masonry in these houses, one face is not enough to establish thickness. Views of both sides of a partition and room context help distinguish genuinely opposing faces from surfaces that merely look close together.

The decision

Faces remain distinct observations linked to the same physical element. Where both are available, their comparison informs the solid. Where one is missing, the gap is retained or any estimate is explicitly declared.

The outcome of this step

Walls described from available evidence, without imposing one uniform thickness throughout the house.

Further reading: Thomson & Boehm, 2015Macher et al., 2017

Distinguish openings and recesses

Does the void pass through the wall or end inside it?

Observations, decision and outcome

Jambs, lintels, sills and backing surfaces help establish the depth of an opening. At Noce, an internal recess does not share the lower boundary of the external opening. At Forno, the recess near the sink has an observed back and must remain distinct from the nearby passage.

The decision

Opening geometry takes priority over leaf and frame detail. Where a jamb is obscured, its boundary remains provisional. The surfaces around an opening are not turned into a complete window product without evidence.

The outcome of this step

Separate passages, openings and recesses, with uncertain outlines identifiable.

Further reading: Gourguechon et al., 2022Patil & Kalantari, 2025

From faces to elements

Which portions can become a coherent solid?

Observations, decision and outcome

Wall portions supported by observations become three-dimensional elements with their openings and irregularity. Successive Noce revisions introduce volumetric walls and the stair-side wall with three openings. At Forno, two internal partitions retain passages and a blind recess.

The decision

The observed surface remains comparable with the reconstructed solid. The boundaries of a studied portion do not automatically become physical wall ends. Hidden materials, build-ups and final quantities are not inferred from shape alone.

The outcome of this step

BIM objects with geometry and identity, retaining the provisional portions.

Further reading: Thomson & Boehm, 2015Macher et al., 2017

Reconstruct the roofs

What roof shape is actually observed?

Observations, decision and outcome

External observations complement the interior survey. Sections compare slopes, edges and changes between roof planes. At Noce, the first trial and the later surface more closely supported by the survey are both retained. At Forno, roofs from the different building volumes progressively enter the assembly.

The decision

The observed roof remains separate from the future roof design. Its surface does not establish hidden structure, thickness or build-up. Gaps in observation also remain visible.

The outcome of this step

Surveyed-state roofs connected to the walls and external observations.

Further reading: Tang et al., 2010

Preserve the actual stair

How does the exterior connect to the Noce upper apartment?

Observations, decision and outcome

The external stair is read as a sequence of individual surfaces. The contribution documents 13 intermediate treads and 14 observed risers, plus the partially observed upper landing surface. The final rise differs from the preceding ones and is not replaced with an average.

The decision

Observed wear and irregularity are retained. The underside, material, hidden supports and railing are not presented as reconstructed. Geometric checks are not a regulatory assessment of the stair.

The outcome of this step

A documented external connection, integrated into Noce revision R006.

Further reading: Gourguechon et al., 2022

Make uncertainty visible

What happens when a measurement is missing?

Observations, decision and outcome

The bedroom–kitchen partition on the Noce upper floor is a concrete example: its two faces cannot be distinguished reliably enough. A provisional estimate of about 80 mm was authorised, while the material remains unknown.

The decision

The model distinguishes the observed face from the estimated one. The assumption follows the element into later revisions and remains a verification request. The same principle applies to obscured jambs, connections and unobserved edges.

The outcome of this step

A model that can support further work without turning a provisional choice into a surveyed fact.

Further reading: Tang et al., 2010Patil & Kalantari, 2025

Check beyond the image

Does a convincing model really contain what it claims?

Observations, decision and outcome

Checks compare surfaces with the survey, examine actual coverage and inspect the exported model. Independent reviews can reject a proposal or restrict acceptance to a portion. Document validity, geometric fidelity and building completeness are separate questions.

The decision

A historical inventory checks that identities do not silently disappear: present geometry, elements without geometry and information awaiting reconstruction remain distinguishable. A numerically faithful conversion does not demonstrate the survey’s physical accuracy.

The outcome of this step

Recorded outcomes and limitations, with rejected trials preserved.

Further reading: Tang et al., 2010Patil & Kalantari, 2025

Build without losing the history

How does the model grow while its steps remain verifiable?

Observations, decision and outcome

Noce revisions add roof surfaces, walls, floor, stair and volumetric portions. Forno revisions extend reconstruction from an initial room to groups of rooms at different levels, then to roofs and partitions. Roof-hidden views reveal what is present inside.

The decision

Each contribution enters the assembly while preserving identities and earlier geometry. Intermediate versions remain available. The 26 September journal records Noce R009 and Forno R006: a milestone in the work, not a declaration of a completed building.

The outcome of this step

Two parallel journeys connected to their case studies and the subsequent shared model.

Further reading: Thomson & Boehm, 2015