すべての動画を見る

Edmund Optics Imaging Lab 2.2: Telecentricity

Telecentricity is an optical principle in lens design that is utilized to guarantee good measurements in gauging applications. Telecentric lenses utilize a unique optical design concept to maintain constant magnification over a specific field of view and working distance in order to minimize perspective error. Join Gregory Hollows, Director of Machine Vision Solutions, as he compares standard imaging lenses to telecentric lenses in three distinct imaging examples in order to illustrate the importance of using the correct imaging optics and optical principles to obtain the best measurements.

Hi, I am Greg Hollows and welcome to the Imaging Lab. This is Telecentricity. Telecentricity is an optical principle in lens design that is utilized to guarantee good measurements in gauging applications. Telecentric lenses are not widely known about outside of the machine vision world but are heavily used for good gauging. It is actually an optical design concept that is different than what most people think it is. In many cases, seeing this for the first-time people wonder if there is a software mechanism going on to maintain these measurements or if there is some sort of zoom property to the lens itself. The telecentric lens, in general, like one of these here, is actually designed to not have any sort of angular field of view associated with it. Most standard lenses will have some sort of angular field of view associated with them and that's why we can see larger objects, objects at further distances away and smaller objects at closer distances. For the telecentric lens, we actually maintain magnification, as the distance either increases or decreases in relation to the lens. This brings about some really neat properties that we can utilize. As one example, these two cubes that we have here. With a standard lens, if I put them at a two different distances away, they'll appear to be different size objects. This is something that we are very used to, like when we drive a car or something like that. We do not see telecentrically. When we're driving a vehicle and we're going down the road, if somebody presses on the brake lights that's two hundred yards in front of us, we know to slow down a little bit. But it is a good distance away, because the car looks very small and our brain is able to understand that it has a distance out. This goes beyond the stereoscopic imaging that we have that helps us handle 3D effects. But a car that's very close to us appears very very large, and we know immediately to to step on the brakes, so we don't crash into it. If our eyes actually saw telecentrically, those cars would look the same size regardless of the distance that they are away. This would be a very bad thing because you would never know when to hit the brakes hard or when to slow down gradually, depending on that situation because the car will look exactly the same size. That is a problem for us. But for an imaging system, it's actually a very important thing. If I was trying again to measure these two cubes that we have here, and even at subtle distances away but I want to measure them to be very very tight accuracy, even that little bit of distance with a standard lens can create enough inaccuracy in the measurement that I would never actually be able to get the repeatability or the reliability that I am looking for. Telecentric lens eliminates this issue because it will see things the same size at those different distances. Another application that we can look at. Looking at this well plate here, as you can see this well plate has significant depth to it and it has a set of different holes in it. With a standard lens, like the one we have on the camera here, as we look down into it, we can see that the bottom of the ones in the middle that are right down at the optical axis of the lens, but as we get to the edges, it is very difficult to see all the way to the bottoms, if not impossible. With a telecentric lens, though, we are actually able to see to the bottom of each one of these wells individually, all at the same time, and be able to pull back the data from that and get information. Very helpful in those sorts of applications. As we can see in the slide here, when we think about how magnification changes, we are going to look at a conventional lens and a telecentric lens, together side by side. What we are going to look at here is a group of pins that are pointing up at the lens, and we are going to see how they are actually imaged. With the conventional lens, you will notice that the pins are actually looking like they are skewed from the center of the image. They are moving further apart, it appears as they get closer, because their magnification is getting higher, and they are getting closer together, the further away they get because the magnification is getting lower in the system. Same way our eyes see things. With a telecentric lens, they actually see things at the same field of view, all the way through this depth that we're looking at. Thus, when we look at them from above, they look like four spots. Another thing that we look at this is when we look at the next slide, you can see some cut-aways of those pins, represented by the individual circles. We can see how the blur patterns, if you are thinking about the ones at the top and the bottom are out of focus, but actually look ovicular when you look at the conventional lens as opposed to being concentric when looking at the telecentric lens. This brings about some neat properties that we are going to look at in a minute and how they can be used even when out of focus. Some other applications as we can see in the slide here. We would be able to actually look at a set of jumper pins that are at different levels and actually, measure the displacement. We will notice in the slide that we have one pin that looks a little oblong in the image and the software information that is wrapped around is oblong. The real-world center and the actual measured center of this object from two different positions because the pin is sticking up towards the camera, it's off to the side and it's imaging at an angle because of this change in magnification in the system. It makes it difficult to determine exactly where it is in real world space and its distance to the pin next to it or on a pin that's at a different level. With a telecentric lens, the real-world position and that actual perceived position by the camera is in the same exact position from center to center, we are able to measure spacing, even on things that are out of focus. This is where we get into some of the depth of field things that go on in the telecentric lenses. Because we have a blur circle that gets created, with a standard lens, it is actually going to spread out as that distance changes, as our object moves like this in the perceived vision of the lens. As things blur, as they go in and out of focus, that blur circle is going to change its position and we can't find the real-world center. The neat thing about a telecentric lens is that since the object is always in the same perceived position no matter where you place it, even if it is out of focus and blurs in and out, the center of that object is still able to be found. So even something that's at a higher position that is out of focus, the center of it can still be found and accurately placed in the system for measurement accuracy. Some other things that telecentric lenses have as a part of them. It's been understood that telecentric lenses have better depth of field than other lenses. This is actually a misnomer. It is actually related to the f/# of the lens and due to the fact that those lenses in general cases have very very high resolution to them and are manufactured to tight tolerances in many cases, actually give you the feeling that you are getting better depth of field than maybe a traditional lens. But it's actually the f/# of the system that drives it and not as much the actual telecentric lens versus a non-telecentric lens. Another thing about telecentric lenses, and here is one of the difficulties that they have. Telecentric lenses can only see objects that are smaller than the actual entrance diameter to the front optic of the lens and actually a little bit smaller than that because we need to accommodate off axis rays for lighting and the fact that we clip certain portions of that when we take a round image circle and place a smaller sensor in the middle of that. So what does that mean? If I want to look, let's say, at this box here telecentrically, I would need a lens that's at least this big in diameter. That's fairly large. Imagine for a second if I said I wanted to get accurate measurements of something like a car door. That is extremely large and would take an enormous lens. First off, it would be very difficult to manufacture it and incredibly expensive and almost impossible to mount very very well because of the size of it and other issues that come into play. You also wouldn't want to hang something like that off the end of the thread to your camera, you could have a real problem. So, what really gets into play there, is telecentric lenses like these can see forty, fifty millimeters of field of view, about two inches. Some of the larger ones that are available on the market go up to maybe four to six inches and the biggest out there that are usually available through custom ordering is somewhere in the twelve to fourteen-inch range. They get very very large and very very expensive. These lenses can be as high as this tall off of the floor in some cases to be able to get the measurement accuracy and the resolution and everything else that you need. They also drive a very high price at those sort of ranges. So that is one of the drawbacks of those lenses. So, if you are going to be using something that needs a high amount of measurement accuracy and you can consider a telecentric lens, something like these are medium sized. When many get so much larger, they have to be accommodated into the system before the rest of the system is built around it. In many cases, when looking for a telecentric lens at the end of the design cycle, the machine has already been built and all the apparatus is in place and there is actually no room to accommodate something of this size. So, if measurement accuracy is key to what you are doing, you need to think about it before hand, in order to be able to accommodate it into your system. That is an overview of Telecentricity. There is more advanced material that will be coming up later on it. Thank you for watching. You can also click on any of the links to take you to another subject of interest.

Edmund Optics Imaging Lab 2.2: Telecentricity

関連製品

製品レビュー