Contact: Rich Kurz   Page content last updated March 20, 2011

Overview

First of all, I think it is pronounced , "hoo'-gin", with "g" like be-gin, and can be found here. I discovered Panorama Tools about 2001, but it was command-line driven back then and more than I wanted to learn. Then this year (2011)I checked on it again, and now it had a GUI and a new name (no longer developed by Professor Dersch) and had become a GNU-licensed application! I discovered it produced better-stitched panoramas than Canon Photostitch, but required more prep. Photostitch does a good job very quickly, and sometimes the results are very good. But often not.
Then while exploring the tutorials I discoverd this one on architectural surveying / perspective restitution. Now THIS  was what I wanted !   Here is what PhotoModeler was doing . . . and for free! !
Does it actually do what it looks like it is doing? Well, that is what these exercises are all about. The short answer is "Yes!" The long answer is it takes practice to figure out how to get there. I share with you my explorations, and hope the results will inspire you to try it as well.

The Explorations

Exercise #1: NCC-1701
Model at the NASM

Exercise #2: A House

Exercise #3: Stereo View

Exercise #4: A Living Room

Exercise #5: Seaview Plans

Exercise #6:Star Trek
Corridor Set

EXERCISE#1--NCC-1701 Model at NASM

TOP LEFT VIEW
I began my exploration of Hugin's features with these photos. I wanted a digital way to derive elevation views. So I started with a simple side eleveation of a personal favorite.
TOP RIGHT VIEW
This trial used some barrel correction, which I now think is not needed unless one used a fisheye lens.
CENTER-LEFT VIEW
Okay. Let's try something more interesting and challenging-the top surface of the primary hull.
CENTER-RIGHT VIEW
Impressive! Note how the engines diverge. They are not on the same plane as the saucer top edge.
LOWER LEFT VIEW
Now for something tough... the underside of the primary hull!
LOWER RIGHT VIEW
Wow! This actually works! It is not quite round, but I think I could get there now with what I have since learned.

EXERCISE#2--A House

LEFT VIEW
The house was built in 1903 as a parsonage to the brick church on the left of the photo. This view shows the right side (relative to the street). The floor plan is nearly square, making all four eleveations nearly symetrical.
CENTER VIEW
This is the right side perspective correction, i.e., in elevation
RIGHT VIEW
The rear elevation, perspective correction. One must ignore the distortions of anything that is not in the same plane as the rear wall!

EXERCISE#3--STEREO VIEW

These are the voyages of the United States Ship Enterprize!
Its three year mission, to show the flag in foreign waters, to protect American trade, and to boldly go and fight the Barbary pirates!

And it is all true! Here is a brief history of the ship. But alas, we know very little about the appearance of that famous namesake from 1800. I found this published photo of a half-hull model purportedly to be that famous ship. So if I can correct for perspective, MAYBE I can also shift the view and make a stereo image!?!? Here is my attempt. Actually although it did produce a stereo effect that looks convincing, I am not at all sure I actually made an accurate stereoview. Remember that corrections are for one plane and one plane only at a time. The rest of the details not in the plane are merely distortions. But decide for yourself.
LEFT VIEW
A scan of the published image (the original is attributed to the US Navy)
RIGHT VIEW
And the resulting stereoview, made by simply shifting the viewpoint to the right.

EXERCISE#4--A Living Room

"Wave House" (Cooper House) / Harry Gesner, architect (1957)
featured in Life, April 16, 1965.  Here is a link to the LIFE article.
TOP LEFT VIEW
This the view published in LIFE. It is a dramatic shot, but noticeably stretched in width.
TOP RIGHT VIEW
The photo was set up to rectify the plane of the floor plan. What I was beginning to realize was that simply correcting for perspective was not yielding correct height/width proportions. Something was wrong. I could stretch it disproportionally in Photoshop if I knew the dimensions... IF was a big word. I spent hours experimenting with the rotation settings on another building, and never figured it out at that time. I tried something different here... I began playing with the focal length and began to realize expected and reasonable results.
I first of all assumed that the central sunken portion was circular. If that is true (and I think it is reasonable), then the entire room turned out to be a square! Interestingly, the focal lenth worked out to be about 20mm giving an angle of view of 90 degrees.
LOWER LEFT VIEW
To check if that was true, the published floorplan, which was shown as a perspective drawing as seen in this view, was rectified.
LOWER RIGHT VIEW
To rectify it in one pass, I first moved the bedrooms down and scaled them as if they were on the same floor level as the living room. The rectified floor plan view indeed showed a square living room. Aligning the room markers visible in the original photo to the rectified floor plan yielded a viewing angle of close to 90 degrees from a position just outside the entry to the living room on the left side!

EXERCISE#5--Seaview Plans

1ST ROW-LEFT VIEW
Here is a screen cap of the fabled plans to the TV submarine Seaview!
1ST ROW RIGHT VIEW
And rectified. Note that the plan is not quite rectangular. I cannot explain why. I tried again and again to set my control points. This has happened often.
2ND ROW LEFT VIEW
This marvelous close-up is another screen cap.
2ND ROW RIGHT VIEW
And rectified. I worked with the first cap to orthorectify it using disportional scaling.
3RD ROW LEFT VIEW
A portion of the fabled plan scanned by Uncle Odie
3RD ROW RIGHT VIEW
The two screen caps aligned to each other. I assumed the 3 cross sections were circular, and kept adjusting the focal length/angle of view until they were effectively circles, as measured in Photoshop. I then fitted the 2nd cap to the 1st, disportionally scaling it until the overlapping details matched, or at least very closely. There were not enough pixels and detail in the 1st cap to be tightly accurate. One way it could be (but wasn't) double checked is to compare the flying sub profile to known photos or drawings.
4TH ROW LEFT VIEW
A photo of the 4-foot model, which closely matches the plan-also by Uncle Odie
4TH ROW RIGHT VIEW
The composite of all the pieces! Again, the first two screen caps were scaled to the model. The scanned plan portion was also scaled to the model. Everything was jiggled to align with each other. AND FITS PRETTY WELL! AND VIOLA--WE HAVE A TO-SCALE SIDE AND PLAN VIEW OF THE MODEL! This is especially useful to establish a somewhat accurate contour of the forward manta fins from a top view.
n.b.  When comparing the size of the midget sub from some drawings I found somewhere (20 ft. length), one discovers that the total length of the Seaview would only be 200 ft !   But in first year canon, the length was stated to be 400 feet, which just happens to match the size of the USS George Washington, the first Polaris sub. And later fan canon established the length to be 600 feet just to be able to squeeze in all the stuff that was supposed to be on board. At 200 feet, the plan shows facilities for maybe a crew of 50 . . . and NO shark tank!

EXERCISE#6--Star Trek Corridor Set

1ST ROW-LEFT VIEW
Here is a screen cap of a typical corridor from the original series, which I think I found HERE.
1ST ROW MIDDLE VIEW
And rectified to the frame opening. I used the blue floor stripe and the red overhead frame for the horizontal alignment, and the wall on either side of the horizontals in about the same positions for the vertical alignment. It looks like a good set of right angle corners!
1ST ROW RIGHT VIEW Once rectified, one can rotate the camera, so to speak. Here is the deck looking straight down, as though in elevation view. I was surprised by the blue and rust red stripes on the deck. I had not noticed them before.
2ND ROW LEFT VIEW
And now, looking straight up at the overhead frames.
2ND ROW RIGHT VIEW
And back to the normal camera position and looking left to the corridor wall. Only the section that the overhead frame and blue deck stripe intersect with is at a right angle to the viewer. That is of course because the other corridor panels angle away as they form the curving corridor. But is one wanted to, one could simple rotate the viewing angle so it is perpendicular to each wall panel. An interesting exercise, though tedious.
Note that I have cropped these images within Hugin.
I have tried this on other sets. I did get a usabe result on the hangar scene from Casablanca, but other scenes can be more difficult if one cannot find good horizontals AND verticals!