Multi-run alignment of survey rounds

Multi-run alignment: seeing change instead of noise

Two survey rounds of the same location rarely produce exactly the same geometry. A different driven line, a different GNSS situation, a different season: there is always a difference that has nothing to do with reality. Overlay such datasets without preparation and mostly what you see is that noise. Multi-run alignment brings the rounds into line first, so what remains really is change: a settlement, an alteration to the road, an object that has moved or gone.

Aligning successive survey rounds on shared control points
Across sensors too, terrestrial alongside mobile for instance
Reporting of the residual, so you know what you are comparing
Every round in the same viewer, ready to be overlaid

Place absolutely first, then refine between rounds

Alignment does not start with the point clouds but with the trajectory. We compute it from the raw GNSS and IMU observations and tie it to surveyed control points, so each round is already absolutely placed in its own right. That resolves most of the difference, particularly the systematic error arising where satellite reception is poor. What remains is a residual difference between rounds, and that is refined geometrically: stable surfaces occurring in both rounds, such as road surface and facades, are fitted to one another, while objects that reasonably move are left out. That way you do not quietly shift the whole cloud to make a parked car line up.

Knowing what you compare before drawing conclusions

An alignment without figures is an assumption. With multi-run alignment we therefore report the residual: how well the rounds fit one another, and where the fit is weaker. That is precisely the information you need to judge whether a difference you find is meaningful. A few centimetres on a stretch where the fit is within a centimetre is a signal; the same few centimetres where the fit is weak is not. For deformation monitoring and periodic inspection that distinction is the entire point, because otherwise capacity goes into chasing measurement noise.

Rounds side by side in the same environment

Once aligned, the rounds sit in GEORIZON in the same frame and the same viewer. You switch between years with a click, overlay point clouds and compare the panorama from then with the one from now at the same viewpoint. That works between sources as well: a terrestrial capture of a structure can sit alongside a mobile survey round and aerial imagery of the same area. For use outside the viewer, connect to the open API or download a selection. Organisations including ProRail (SpoorInBeeld), the municipality of Nijmegen and Kempkes Landmeten work with their geodata this way.