Speaking of my own collection, here's another tool I've created for my Test Model
. I call it that, for it's my working model until I get my drawings completed. It is based on BJ West's
model, which was based on Simon's drawings. While it is unsuitable for accurate measurements, it is suitable for zeroing in on those measurements, as a tool to help me analyze some of these photos.
For instance, using modo, and some information I have on the original model, I was able to create a bounding box around my test model, that is, the proportional equivalent of the original model. My data holds that the original was a scale model for an actual ship that would have been 175 feet long, and 85 feet wide at the wingtips. Now, I don't know how Simon came to the conclusions for his drawings, as to size or scale, but if I superimpose a grid scaled to 175 feet for length, the width of the wingspan in his drawing comes out as 88-89 feet, and this was followed pretty closely by BJW. I don't know if he had the best evidence for his wing calculations, or the length, but using the length as the standard, I will accept as a working hypothesis that his height is close to correct, which is 19.4 feet. This figure is only moderately reliable, due to several factors. Simon drew the bottom of the wings on the same level as the bottom of the ship, and that is how BJW modeled it. I will have to scale the size of the wings down, width-wise, and height-wise, to fit the rest of those features and make the modeling changes at the bottom curvature of the ship. As you can see, it's a hairy problem, but one I believe I can unravel.
By wrapping a semi transparent bounding box around my test model, that is divided up into 1 foot squares, and level with the highest point on Orion's Engine Bell Housing, and the length of the fuselage without the probes, I can compare individual photos of the original model, as to the angle of attack of the ship, and the type of lens used to take the photo, and distance from the model (part of the data I have and am gathering), ie., the focal length. Then I can do a test render by first orienting my camera in modo to a close approximation of the photo, and compare them. Each adjustment will get me closer to two images that have an identical silhouette. Differences that are due to lens distortion can be compensated for. Then with a render of the bounding box, I can draw an accurate perspective drawing with a horizon and vanishing points, and thus plot the location of all the visible features, which I can save to yet another drawing, as I do each image, I will add this data to that final data image. Comparing the data from all of the photos will give me a pretty decent idea of where everything goes, their proportions and distances.
With that data, I will be able to create orthographic drawings of a great degree of accuracy. Sort of a graphics version of reverse engineering
Here's my first render of the test model, with bounding box. You may be able to notice the wingtips protruding from the sides of the box. I'll get around to fixing that in it's own time.
I'd appreciate any crits on my concepts for a solution to this problem, or tips on enhancing any of them, or especially, any better ideas, tools, features of modo that would help make this easier, or even about any scripts you may know about that help making drawings of this Cad type operation.
Message edited by LeonArtO3D on 11/23/2010 - 2:07 AM