PDA

View Full Version : Depth Perception



K.L. Davis
06-06-11, 10:38
Humans can be pretty good at estimating distance... some are much better than others, but nearly everyone can improve their ability to estimate distance. Billed correctly, we are talking about depth perception... and it is quite interesting how many ways we go about it...

One of my rats has a habit of swaying from side to side when she is looking at you; it looks almost like she is drunk... and it kind of creeps some people out. What is really going on is that rats have very poor eyesight and she is using motion parallax to determine the shape, size and distance of what she is looking at - you can see a lot of animals (even humans) do it, usually with a cocking of the head, leaning slightly, or taking small steps to one side, then the other.

Still, it does look a little like she is sizing you up as a snack.

So this motion parallax is one of the methods that we use to determine depth or distance... there are others of course, but before we get into them, you have to understand something about human vision: it is complex to the point of being miraculous, and faulty to the point of being unreliable - all in all though, amazing.

Anyone that has worked in law enforcement can tell you that an "eye witness" may not be much help at all... for a large part, we see what we "think" we see, and all too often we only see what we want to see. Check out this video for an example...

http://www.youtube.com/watch?v=vJG698U2Mvo&feature=related

So, here is the deal: our eyes can only see a two-dimensional image, and each eye is independent... the image from each eye is sent to the brian, where they are combined, interperated and perceived as a single three-dimensional image. This image is called the "mind's eye" or cyclopian image. Often our mind takes some liberties with forming this image, combining objects, filling in missing parts of the field, adding colors or patterns, and even eliminating things. Cool, huh?

When talking about depth perception, one thing that you will often hear is that we use the distance between our eyes, to "triangulate" the distance to an object - simply put, that when we focus on an object, our eyes have to turn in a little to look at the object, and we use the angle of our eyes, to determine the distance to the object we are looking at.

Not entirely. We do this to some extent - it is called convergance - but the distance between our eyes is really not that great... it is only about two and half inches for an adult male. The fact is, we use this method with objects that are close to us, usually within arms length and never beyond about 10 feet or so.

So, some of the other techniques that we use to determine distance:

Motion Parallax - We already talked about this a little... take a moment to check this out though. When we look across a distance and move laterally (sway back and forth like my rat), objects in our field of view will be displaced or appear to move opposite the direction that we are moving - the closer an object is to you, the further and faster it will appear to move.

Interposition - This is simple, if something is blocking the view of another object, we assume that the object being blocked is behind the other object.

Linear Perspective - Think of this as how a road gets narrower, as it gets further away.

Angles - We have expectations of angles, especially man made angles... we expect a wheel to be round and a window to be square; our expectation of angles and shapes is often used to estimate an object's size or distance. This is the reason that the well known "Ames Room" illusions work so well.

Relative Size - When we are looking at two or more of the same size objects, we assume that the ones that appear smaller are further away than that ones that appear larger.

Light and Shadow - The location of light source and shadows add depth to our view... and an object being lighter or darker can make it appear closer or further away.

Motion Blur - When we are moving, objects close to us will be blurred by that motion, the closer they are, the more they are blurred.

Texture - The closer an object is, the more texture feature we will see... standing on a gravel road, you can make out that it is gravel for several yards, then it just appears rough, and further out even that detail disappears.

So why bother will all of this? Well, this is what is going on in your brain-housing group anyway, might as well have some understanding of how it works; this will actually help you learn and improve your ability to estimate range.

Of course the bottom line is that you have to get out and try it... try to spend some time now and then working on estimating and verifying distances - you might be surprised how good you can be at it.

========================
Originally posted on my facebook page (http://www.facebook.com/DMRSurvival)

Failure2Stop
06-06-11, 10:47
Interesting topic for discussion, thanks.

uwe1
06-06-11, 16:08
Kino, great write up and in a very easy to understand format.

I don't know what reference materials you have, or if you've already got this, but this book was one of the ones I was assigned to read during my first year:

http://www.amazon.com/Visual-Perception-Orientation-Steven-Schwartz/dp/0071411879

It has a lot of useful information about how and why we see what we see.

ReaperAZ
06-06-11, 16:28
Great post Mr. Davis! A friend of mine growing up had a rat that did the exact thing you speak of. It was kind of strange to watch it do it the first few times, but after a few times I just ignored it. Really makes me think about how the guys "back in the day" did their range estimations and such without all the fancy lasers and range finders we have now.

Iraq Ninja
06-06-11, 23:47
Interesting...

The first thing I thought about was NVGs and the problems one can have with depth perception.