Estimate object size

Question

If I have a [digital] photograph of an object, can I estimate it's approximate size?

I notice with interest that my camera (Nikon D3300) seems to record the approximate focal length of the lens in the Exif data. (It seems to round it to the nearest "common" nominal length.) However, I see nothing in the Exif data about the plane of focus. Presumably the camera system has that data somewhere (e.g., some lenses have it printed on the focusing scale), but I don't see it recorded anywhere.

Assuming I had the lens focal length and the lens focus distance, presumably it should be quite straight forward to estimate object size.

(In case it matters, I'm particularly interested in extremely tiny objects such as ants, grains of sand... stuff you can't exactly take a ruler to. I don't know whether that makes size estimates easier or harder.)

Answer 1

I am told that roofing estimates use satellite photos to measure roof size. But their distance is more than a few feet, and their focal length is known, and of course, they have one situation which they can learn to correct for.

But for your purposes, No. For zoom lenses, the focal distance in the Exif is far wrong. For most zooms, it will indicate a radically different distance numbers for every zoom situation, not even close. Do some measuring and checking for yourself before believing the numbers you see. Measure the actual distance and then include zooms at both ends of its range.

An article about this, about how inaccurate lens distance hurts Nikon TTL BL flash exposures is at http://www.scantips.com/lights/ttlbl-d.html There is a chart there of inaccuracy of a few lenses.

The nominal focal length is in the Exif, but the marked focal length is applicable to infinity focus. At close distances, 2 or 3 feet, and certainly for macro situations, the focal length will be much different than marked. In fact, the definition of 1:1 macro means that the object focus distance is equal to the lens focal length (necessarily equal distances in front of and behind the lens at 1:1).

That said, there is a calculator at http://www.scantips.com/lights/subjectdistance.html that if object size is known, and if distance is several feet or more, it computes distance from object size, given a object distance greater than a few feet. Trial and error could work the other way, for object size. The math is shown there.

Answer 2

You can trace a triangle from the image of an object back to the lens (inside the camera). The height of this triangle is the focal length. The length of the image of the object is the base of the triangle. Suppose the focal length is 50mm and the length of the image on film or sensor is 8mm. We find the ratio thus: 50 ÷ 8 = 6.25.

We can also trace out a tringle from the object to the lens. The Exif file gives both focal length and object distance. Suppose the object distance is 600mm. We divide thus: 600 ÷ 6.25 = 96. We have calculated the length of the object to be 96mm.

The caveat is: we need to measure the image size. If we were dealing with film, we can simply measure the image using a millimeter rule. For the digital camera we can display an image on the computer screen and make the measurements. The image on the screen must be the same size as the format. For the FX, that’s 24mm height by 36mm length. For the DX that’s 16mm height by 24mm length. You can consult your camera specifications to get the frame size. Also, you need to control the image size to take into account the magnification, if any, that your display software applies.

The principle is: The object height to lens traces out a triangle. The image height to lens traces out a similar triangle. Simple ratio math is used to calculate the unknown object’s height provided the object’s distance is known as well as the focal length.