Hull lines from existing mold

I have been asked to produce a set of lines from an existing mold lying in someones back yard. I am probably going to use Maxsurf and/or Rhino to do the lines, BUT what hardware is used to scan and take the measurements accurately from the hull surface. This is a motor cat of about 42 ft on the waterline, so not something you can try and measure up by hand.

I would appreciate links to anyone selling these "scanners", or any other suggestions you guys may have.

Thanks - Gustav

 

You will be better off hiring a company that specializes in that. Buying the equipment and software, then learning how to use it won't be economic for a single job.

 

If you get it scanned as an IGES points file, you will get a LOT of points...

Best plan with one of these is to remodel it, using selected sections from the points file and some real measurements. Compare the result with the original file, which will be slow to render because you probably have 60,000+ points. Maybe even 200,000+, trouble is if you simplify too much you may loose some data on a complex part. A bit easier if the hull is symmetrical though when you only have to deal with one half, but this assumes the moulds are fair and symmetrical too and they may not be.

There is some software which uses specific photographs, you may be able to generate a reasonable hull/mould from. Search this Forum for that specific use.

Rhino should have enough tools to get a decent surface(s). Rather depends too on if you want to CNC other parts to fit etc etc or just for visualising.

 

3D measurement software

Hi gonzo and Sukisolo,

I've been to see a chap out here who sells/supports Photomodeler. This s/w (and probably the others too) is more then capable of taking 3D measurements off an existing hull or mold. There are quite a few intricate tricks to the process, but no “rocket science” involved. A reasonable camera with a fixed focus lense (preferably) and the s/w is all you need to take measurements. I'm told accuracy is 0.2mm per 10m, which I think is roughly 1/16” per 30ft. That's pretty good by my reckoning. From there you export the point cloud or surface into your modelling s/w and develop from there. It looks like I will be out of pocket to the tune of $1200 excl. camera, but this has downstream applications and I could probably earn that money back in time.

I think I should add a disclaimer here! The folk selling the s/w like to tell you its not difficult to learn and the process is reasonably simple... which of course means it probably isn't... or else they wouldn't have mentioned this. Expect a 2 – 6 week learning curve, and practice on smaller objects before tackling the big project.

Here are links to the s/w vendors, and yes Google are getting involved too... there goes the farm!!!!


http://www.agisoft.com/

http://www.photomodeler.com/index.html

http://www.rhinophoto3d.com/index.php

https://www.google.com/atap/project-tango/

 

Sounds like you have a solution. Also the accuracy sounds extremely good, more than enough to get a good model from. Probably fairer than the original hull....

My own take would be to still take physical measurements on any critical areas, to double check and ensure a good fit, if engineering something onto the hull. Maybe that is over cautious but it is cheap insurance!.

Good luck with it.

 

The price and accuracy sound really good. Can they give you a demonstration? I think that taking physical measurements to a few points and comparing them to the results would be an easy way to verify the accuracy.

 

The man who sells/supports Photomodeler will die laughing if you intend to do this verification. You can reject the offer as being too expensive or for any other reason but never for not being too accurate. And above all, compare the accuracy manually with a tape or something similar, it seems unsuitable.

 

I have used PhotoModeler to measure several boats including an 80+ foot sardine carrier and a launch on a trailer. PhotoModeler usually requires targets to be applied to the surface. My current preference if the surface has sufficient "visual texture" is to use PhotoScan which does not require targets and generates a cloud of points. A clean mold will not have sufficient visual texture for PhotoScan and will require applied targets. If the mold is dirty and the dirt is not uniform then PhotoScan may work.

The first boat I measured with PhotoModeler was the Jacob Pike, an 80+ foot sardine carrier. For this one I used a small number of applied targets and a network of applied tape on one side of the boat. Over 1400 3D points were generated using PhotoModeler. Most of the points used by the software were manually identified at the intersections of the tape and the tape and planking seams, and this was very time consuming. (Subsequent projects used applied targets which can be automatically recognized and located by the software.) The points were imported into Rhino3D, surfaces created through the points, and then lines generated by contouring the surfaces.

Another boat I measured was the Lindy Lou, a 30 foot fiberglass launch on a trailer. Applied targets were used for this boat covering both sides and the bottom. (Tape lines were used to guide placement of the targets.) Access for photos was impeded by the trailer, and photos of the bottom of the boat were taken laying under the boat with the camera against the ground. 1669 3D points covering both sides and the bottom of the hull were generated using PhotoModeler. The points were imported into Rhino3D, surfaces created through the points, and then lines generated by contouring the surfaces.

A zoom lens works fine with PhotoModeler as long as the zoom setting remains the same. I sometimes use a small piece of tape to keep from moving the zoom ring.

Accuracy of 0.2mm per 10m (0.008 inches in 33 feet or 0.002%) is optimistic.

View attachment Jacob Pike T101 Aft Qtr 2.pdf