Jet Field Notes

Practical LiDAR workflows in the field

Independent publication · No official affiliation

Field Notes / Application guide

From Jet Point Cloud to CAD or BIM: A Deliverable-First Workflow

A lean point-cloud modeling workflow for coordinates, source provenance, extraction rules, IFC exchange, layered QA, and a usable final package.


Before-you-go brief

Use when

The environment, access plan, and requested output match the workflow described in this note.

Verify first

Confirm project coordinates, acceptance checks, site restrictions, and the required evidence of coverage.

Bring or prepare

Carry the approved control plan, route map, safety plan, field log, and redundant storage appropriate to the site.

Stop or escalate

Pause when access, control, coverage, safety, or deliverable assumptions no longer match the approved plan.

Workflow route

On this route: 01 Scope · 02 Capture · 03 Process · 04 Handoff

The fastest route from a Jet point cloud to CAD or BIM starts with the deliverable. A team that opens a dense cloud and begins tracing everything will spend time on geometry nobody needs, invent detail in occluded areas, and struggle with coordinates at handoff. A deliverable-first workflow limits the model, preserves source provenance, and directs QA toward the features that drive decisions.

Artec currently lists Jet outputs including LAS, LAZ, PLY, E57, full-resolution and decimated point clouds, and a trajectory file. Those exports can feed several processing and authoring tools. Software choice should follow the recipient’s formats, coordinate behavior, classification or segmentation needs, automation, licensing, hardware, collaboration, and archive policy.

Create a model definition sheet

The model definition sheet names purpose, coordinate frame, units, extents, element classes, level of geometric representation, attributes, tolerance, exclusions, native format, exchange format, sheets or views, and acceptance test. It should distinguish visible as-found surfaces from nominal or design-intent objects.

Deliverable Model only what supports Primary QA view
2D as-built CAD Required plans, sections, breaklines, assets Sheet-based overlay and spot dimensions
Coordination BIM Clash-relevant envelopes and systems Sections through congested zones
Facility model Maintainable assets, spaces, access context Asset ID and location review
Surface/terrain Ground, hardscape, breaklines, exclusions Cross-sections and checkpoints
Digital twin context Approved geometry linked to stable IDs Source date, coordinates, and provenance

Model detail should be asymmetric. A project may need precise openings and equipment clearances while accepting simplified walls and generic furniture. That targeted scope is more useful than a uniformly elaborate model.

Protect the source chain

Keep raw observations read-only. Create a registered master in the approved project coordinate frame, then derive clipped, decimated, classified, or colorized working files. Every derivative receives a source reference, processing version, date, coordinate statement, and resolution or sampling note.

Record control used to place the cloud and independent checkpoints used to test it. A model cannot be more defensible than its source. If the point cloud has accepted limitations in one wing or level, carry those limitations into the model work plan.

Test coordinates before loading everything

Large real-world coordinates, local origins, rotations, vertical datums, and unit conversion are common failure points. Export a small representative zone with two or more known points. Load it into the target CAD or BIM environment and verify coordinates, elevation, orientation, units, and shared-coordinate behavior.

If a temporary working origin is necessary, use a documented reversible transform. Deliver the transform and test it in both directions. Never rely on dragging a cloud visually into place.

Prepare clouds by model zone

Split working files along meaningful boundaries such as floor, building wing, process area, alignment station, or civil tile. Include controlled overlap at boundaries so elements crossing zones can be checked. Preserve a master index showing each file’s extent, revision, point spacing or decimation, and capture date.

Filtering should remove obvious transient noise without erasing thin required features. Classification can speed terrain or asset work, while automated results require review. Decimation should be tested on the smallest important feature. A lightweight file that drops every cable tray edge has failed its purpose.

Extract with explicit rules

CAD and BIM extraction rules describe how observed surfaces become deliverable objects. For a wall, specify whether the line follows one observed face, a fitted centerline, or a nominal assembly. For a pipe, define centerline fitting, diameter source, insulation treatment, and minimum observed arc. For a floor, define filtering, breaklines, grid or triangulation, and treatment of openings.

  • Observed: geometry directly supported by approved point-cloud surfaces.
  • Inferred: geometry completed from partial observation under a documented rule.
  • Referenced: geometry or attributes supplied by owner records or design documents.
  • Excluded: element outside scope or without sufficient evidence.

Store provenance through parameters, notes, layers, properties, or an external issue register appropriate to the software. The recipient should be able to tell what was measured and what was assumed.

Use open exchange carefully

buildingSMART describes IFC as an open international standard for sharing BIM data. The current official IFC documentation includes schema, property and quantity definitions, and exchange mechanisms. “Deliver IFC” remains incomplete unless the contract names the version, model view or exchange requirement, elements, properties, coordinates, and validation method.

For point-cloud handoff, test LAS, LAZ, or E57 imports in the actual software path. Check coordinate precision, color and intensity mapping, classification, compression, and scale. A format supported on a feature list may behave differently at production size.

QA the model in layers

  1. Coordinate QA: verify project points, units, rotation, and elevation.
  2. Coverage QA: confirm every scoped zone and element class is represented or logged as excluded.
  3. Geometric QA: compare samples through overlays, sections, dimensions, or surface distances using the agreed method.
  4. Semantic QA: check names, systems, levels, asset IDs, properties, and classification.
  5. Exchange QA: open native and exchange files in receiving software and inspect warnings.
  6. Documentation QA: reconcile manifest, revisions, source dates, limitations, and issue register.

Sampling should cover all zones and difficult feature types, with extra checks at joins, stairs, shafts, racks, and inferred geometry. Keep the modeler separate from at least part of the QA review.

The model is a controlled interpretation of observed geometry, with every assumption visible enough to review.

Deliver a usable package

The package contains accepted point-cloud sources, native CAD or BIM, agreed exchange files, sheets or views, coordinate statement, control and checkpoint summary, model definition sheet, extraction rules, issue and exclusion register, software versions, file manifest, and correction contact.

A deliverable-first workflow feels lean because it is lean. The team spends attention on the pipe that drives a tie-in, the wall that controls clearance, or the surface that sets drainage. That focus turns an enormous point cloud into a model people can actually trust and maintain.

QA and handoff gate

  • ✓ Compare the registered dataset with the approved control and record exceptions.
  • ✓ Review occlusions and coverage gaps against the requested use.
  • ✓ Inspect registration joins, loop closures, and any segmented runs.
  • ✓ Confirm coordinate reference, project datum, units, orientation, and file naming.
  • ✓ Package source records, processed files, field logs, and known limitations.
  • ✓ Obtain stakeholder acceptance against written criteria—not an assumed visual check.

Adjacent field routes

Set the field control plan

Use this first when coordinates and independent checks still need agreement.

Carry capture into deliverables

Continue here when processing and acceptance packaging are the next decisions.

Choose another deployment mode

Compare access and risk conditions before changing the capture platform.

Methodology · Request a correction