Batbrat
Member
- Joined
- May 5, 2016
- Messages
- 358
- Reaction score
- 3
Working with your own image, lines and measurements, I arrive at a suspect height of 5'5" to 5'8".
I think that the distortion is the vertical millimeter bar - it must be parallel to and angled the same as the 'vertical' angled green suspect line and also the 'vertical' angled red center 84" door line to come out correctly? because this is a one point perspective prism?
I put the photo fullscreen on my computer and measured your lines with a ruler in inches. The suspect is presumed to be standing at the center of the door and close to it.
FIRST estimate: The 'vertical' angled green suspect line = 3.25 my ruler inches, and the 'vertical' angled red center door line = 4.0 my ruler inches. If the red center line is actually 84", then the math is: 4/3.25 = 84/X, solve for X with 4X=84*3.25, then X = 273/4 = 68.25 actual suspect inches. Allowing 1" for boot heel and 2" for helmet top, 68.25" minus 3" = 65.25" or about 5'5" suspect.
SECOND estimate (same process, trying to measure more accurately with my ruler against CRT screen with image larger): The 'vertical' angled green suspect line = 4.75 my ruler inches, and the 'vertical' angled red center door line = 5.625 my ruler inches. If the red center line is actually 84", then the math is: 5.625/4.75 = 84/X, solve for X with 5.625X=84*4.75, then X = 399/5.625 = 70.9 actual suspect inches. Allowing 1" for boot heel and 2" for helmet top, rounding 70.9 up to 71 inches, minus 3" = 68" or about 5'8" suspect.
What would happen if your MM bar were placed between the suspect and the door, and also angled to be parallel with the 'vertical' angled green suspect line and the 'vertical' angled red center door line? (eg, the 'vertical' angled green suspect line is longer than the MM vertical line as shown, distorting suspect height to taller. ???
Adding: My math was trying to compare the red center door line to the green suspect line, since as you pointed out for us, the fish eye lens is very distorting making objects in the center (MM bar) different than (slanted) objects to either side (suspect and door) ??? I don't know if this makes any sense!!! LOL The MM bar would have to stick out at the top from the screen toward the viewer to be in the same plane as the suspect and the door.
Yes your conversion math is correct. Since it's difficult to determine if the distortion of items in the image is caused by the top-down perspective or the wrap of the wide-angle lens, I have no way of knowing if a distortion needs to be corrected or illustrated. The image is cropped, but in the original frame both the subject and the door are near the center, so they're arguably distorted in comparable amounts. That's why I chose that frame to isolate.
The ruler is simply a scale, and since the quality of the footage is poor, a single pixel could mean as much as half an inch in variation. It is by no means "accurate", but it's the best I'm able to do given the image quality. However, at your suggestion, I've included an example scale to the right in the image below. You might be surprised to notice that if you put a scale ruler on a simulated top-down perspective, it means the subject is even taller than my estimates (see A below) over 6'1", and LE's estimates of 5'2'-5'6" slide further down the scale ruler (see B and C below) because of perspective. So more discrepancy (between my estimates and LE's), not less.