I am preparing a tender for the survey of a Roman site so to see if I can use a scanner and VirtuSurv to prep 1:200 scale plans I got the loan of both for a day. The FARO Focus 3D scanner is small, light and easy to use.
For a change the weather was dry and eventually there was enough wind to get a camera aloft so I was able to capture both photo and scan data to work with.
The image resolution is fine for 1:200 so with sufficient control I could trace off the images to get the plan. So far so good. I extracted 3D data from the scans in VirtuSurv:
I discovered VirtuSurv works best if you work on one or two scan views at a time, it’s easy to open and close the views as you go. The PacMan view at the bottom left shows me where the scan is and the direction of the view from the scan centre. I can select the next scan I need by picking the next PacMan and the scan is selected in the project tree from where I can open the RGB view.
On the left the digitised lines in CAD, on the right the VirtuSurv scene of one of 5 scans captured showing the points I clicked to get the plan: this is a very big deal- I’m able to pan and zoom around a colour image and get 3D points spat out into CAD. I have not had to look at a point cloud at all, no meshing or modelling is needed, no point clouds have been bled into AutoCAD… and I’m able to trace lines from the scan data. Amen. Job done. There is no better way. End of story.
So at last I have got a bomb proof scan to plan route. The scans need to be registered in Faro Scene but once that’s done I never have to look at the damn dots again. With the trace as a control I set up photo-control points in CAD, this was a bit tricky as there is a cross- fall on the site so finding a logical reference plane to project the control points onto was a matter of judgement. Using PhotoPlan I was able to rectify the KAP image and the tracing can really begin!
In this case the best image by far to trace is still the nadir KAP view, working with the scan data in VirtuSurv view is pretty straight forward and much quicker than working off the dots in CAD but as the scans were quite oblique to the wall tops the trace was slower than from a single image for the same area of cover. Certainly the scale tolerance of data traced off the image is going to be worse than from the scans but the economy of cover is greater; once again I find a combination of techniques is needed to get an optimum result.
The good news is I can get photo control points easily into CAD for PhotoPlan from the scans once they have been loaded into VirtuSurv.
More on VirtuSurv here