How to Create Lines From A Simple Drawing

David,
the software just needs to identify the same points from one pic to the next.. I don't think there is any requirement about the pattern being random there just need to be enough visually distinct points to define the surface. Did you try the marker method? What were the problems?

 

Instead of an erasable marker, maybe colored circular stickers. Might photograph better.

 

Maybe. I think in terms of cross hairs. The software is trying to place pixels from one pic to the next.

I looked some more and there is still a free Autocad program, iOS, PC, and cloud.

http://www.123dapp.com/catch

I will get it on my iPad and try scanning my kayak.

 

Reverse engineering is an exciting methodology and 3D scanners are impressive.But with all due respect I think that in the present case, it would be like to catch flies with cannon.

 

I have tried marking a hull, tape, and colored circular stickers. None work reliably with the SfM type of photogrammetry such as AutoCAD's 123Catch, PhotoScan or PhotoModeler Scanner which automatically identify and match points. The surface needs a random texture with a small enough scale and dense enough distribution that a sufficient number of unique points are identified and matched between photos by the software. A clean, plain surface or even a repeated pattern does not work with this type of photogrammetry. Random scribbling on the surface or apply relatively large scale features such as dots or tape isn't sufficient. For surfaces with a good random visual texture this type of photogrammetry can work extremely well.

There is another version of close range photogrammetry used in PhotoModeler and RhinoPhoto which can use applied targets. PhotoModeler allows both manual marking of points and photos and matching of points between photos as was as automating the point marking process by using dots or coded targets, and automating the point matching process with coded targets. RhinoPhoto works only with coded targets. I've done several boats using applied targets of various types and PhotoModeler. Significant desk work is needed after taking the photos. A while back I sent a set of coded targets to a boat owner who was very knowledgeable about taking lines. I included instructions on where to apply the targets and what digital photos to take. He did so and sent me the photos. I processed the photos and developed a 3D virtual model and a lines drawing.

To create a set of lines for a display model where profile and deck drawings already exist it may be simpler and more efficient for Southern Cross to measure a few sections along with some photos and have someone with experience such as TANSL create a set of lines from that set of information. That of course depends on Southern Cross's ability to take the measurements and photos. If the boat was readily available to me and I was going to create the lines it would probably be quicker and simpler for me to apply some targets, take some photos and process the photos in PhotoModeler to obtain a set of discrete 3D points on the surface.

 

Tansl has offered to help me and has sent me a diagram of the specific data points he needs. I'm pretty meticulous about these things so I expect my measurements to be somewhat accurate.

I'm excited about this. The Olson 30 is an historic race boat in these parts, one of the original Santa Cruz ULDB's. A true thoroughbred ocean racer. It's still a "giant killer" in some conditions. There has never been a hull plan to refer to. Hopefully, we can create one. I'll post our progress (with Tansl's permission of course).

 

Sounds good!

 

Some inside views of an Olsen 30 for reference here

http://www.fiberglasssupply.com/pdf/vacuum/CONTOURKORE.pdf

 

Mr. Watson,

Thanks for that. In all my searches I have never come across this. Side note - I grew up in Watson's Bay, Sydney.

 

A simple home builders laser level should make pulling off the lines easier.

 

It was sheer luck ( or should I say inner hull and sheer luck ) in the search engines. Isn't the internet wonderful - sometimes !.



Unfortunately, I dont think i am related to anyone with interest in prime real estate so close to a capital city, though i would consider adoption

 

Just caught this. Obviously the wrong link.

I bought a Bosch Laser. Accurate to .001M. Will mount on a Slik tripod. Setting up an accurate perimeter should take some time. Leveling the boat will take some time too. And each measurement, 50+ should just about do me in.

Mr Watson, passed the article you found through the community. No one had ever seen it. Great find. What search engine do you use?

 

Which model Bosch laser?

 

At first I thought that you would need a multi-scanning laser, to create a point cloud for the Cad. However, on further reading, that would have been way overkill.

Dont forget though, that millimetre accuracy is not required. .001 mm is not needed - +- 1 mm is plenty. The hull design may be good, but no hull was ever built to 1 mm precision. You will only need to do a half hull, so symmetry isnt a problem as Cad will copy the other side perfectly, and any fairing issues can be solved in the Cad conversion. .

Truth be told, even a laser may be overkill. You could simply run a metal rail parallel with the keel, and take 8 or 9 station offsets manually, and have a great result. Its not like the wing of a jumbo jet where every mm is critical, thank goodness.

Dont sweat the job - it doesnt need rocket science. have fun.

 

I've found the cross-line laser level is a huge time saver.

Like this https://www.boschtools.com/Products/Tools/Pages/BoschProductDetail.aspx?pid=GLL2-50

For the topsides you set it up on a vertical pole forward and aft just beyond max beam. The horizontal line is a waterline on the hull, the vertical is your base (reference). Measure the distance between the two at common interval (stations). Do the opposite under the boat, horizontal line is reference, vertical is a buttock line on the hull.