Conversion

Orthophoto to CAD: from photogrammetric output to DXF

Orthophotos are metrically correct but still raster. This is the workflow for survey offices that need consistent measurements across the whole surface.

Photogrammetric orthophoto of a historic elevation used as the source for CAD vectorisation

Source photograph

DXF vector CAD drawing produced from the orthophoto, with openings, mortar courses and cornices as real-scale polylines

Vector output (SVG / DXF)

A different user: the survey office

Readers of this page have already done the metric work: a photogrammetric campaign, a laser scan, a rectification in Metashape, RealityCapture, PhotoModeler or an equivalent pipeline. The deliverable in hand is an orthophoto or rectified photoplan — a raster with constant scale and perspective removed. What remains is the step that dominates the budget: turning that raster into CAD geometry.

Compared with the single-photograph route, the accuracy regime changes entirely. On a rectified source, two-point calibration no longer interprets perspective; it only fixes the scale factor. Measurements stay consistent in every direction and across the full extent of the surface — the condition heritage and infrastructure specifications require when a drawing must be spot-checked.

What gets extracted on large elevations

The engine vectorises opening outlines, mortar courses and joints, springing lines, jambs and lintels, stone elements, cornices and string courses, together with losses, detachments and structural cracks. On frescoed or mosaic surfaces it follows the boundaries of colour fields and of the larger tesserae. The output is a base drawing for as-built elevations, thematic decay mapping and surface quantity take-off.

Orthophotos are usually heavy files, and the input accepts JPG, PNG and TIFF up to 40 MB. In practice, export at the lowest resolution that still resolves the detail you intend to vectorise: the ratio between pixel size and real size determines how fine the extracted paths can be.

Spatial coordinates and georeferenced insertion

Before generation you can declare the drawing's spatial coordinates: insertion X, Y and Z, real width and height, rotation, and origin at the lower-left corner or at the centre. The DXF therefore lands directly in the project or GIS coordinate system, with no manual repositioning to introduce error.

The DXF imports into AutoCAD, BricsCAD, Rhino, QGIS and QField, where it can be overlaid on point clouds, laser-scanner sections and existing survey layers. The SVG covers layout, presentation plates and attachments for authorisation procedures.

At a glance

  • Input: rectified photoplans and photogrammetric orthophotos, JPG, PNG or TIFF up to 40 MB
  • Consistent measurements across the surface, not only along the calibration direction
  • Optional spatial coordinates: X, Y, Z, real size, rotation, origin
  • DXF ready for AutoCAD, BricsCAD, Rhino, QGIS and QField
  • Base for as-built elevations, decay mapping and quantity take-off

Questions

What export resolution should I use for the orthophoto?

Start from real pixel size: if the finest detail you must vectorise is a few millimetres wide, you need two or three pixels across it. Exporting beyond that only increases file weight without adding usable geometry.

Can the drawing be overlaid on laser-scanner point clouds?

Yes. Setting insertion point, real size, rotation and origin in the spatial-coordinates section aligns the DXF with the cloud or with sections already in the project, with no manual repositioning.

Is it acceptable for specifications requiring verifiable deliverables?

On a rectified source, yes: the scale is declared, the file is a DXF of real entities, and the calibration measurement is documented. Spot-check verification and final validation stay with the professional signing the plate.

Other conversions

Prova CVCAD sul tuo prossimo rilievo

Registrati e ricevi 1 conversione gratuita: carica una foto e scarica un DXF già in scala millimetrica.

Apri l'app