Angle of View

[BUGSBUNNYBOSS]

I understand that, smaller the angle of view, greater the reach of the lens (correct me if i'm wrong).
Now I want to know how is the distance (reach of the lens) associated with this angle. I'm considering my subject as a bus or a car.

What is the difference between lenses whose angle of view is 1 deg. and 0.5 deg. in terms of their reach.

I am guessing the reach would be double for 0.5 deg. lens. Can someone give a numerical value of the distance and tell me how do I calculate the distance?

I was browsing Flir systems website and noted that, for one of their camera's whose angle of view is 0.25 deg. the reach was 9km (saw it in one of their demo video). Is this value a constant for all the lenses with 0.25 deg angle of view ?

Finally, What are the factors affecting the angle of view? and is there any relation between it and picture quality?

[davidedric]

Back to schoolboy trigonometry, I think.

To me, it doesn't make sense to say a particular lens has a certain fixed reach. What we are saying is how big an object will fill the frame at a given distance.

The formula would be size of object = tan (angle of view) * distance to object

The tan function is not linear, but at small angles and distances it won't be far our, I. E. 2 degrees will require an about twice as big, at a given distance as 1 degree.

Hope my memory of 50 years ago is right!

Dave

[BUGSBUNNYBOSS]

I did some reading and found this link. please refer
http://www.gd-imaging.com/Documents/...DS5-213-10.pdf

Can someone explain the terms in visible camera section. what does 440kpixels in (440k Pixels; 752(H) x 582(V)) mean?

Will it make sense to compare this sensor, with a full frame sensor. If yes, how will be its performance comparing a 35mm sensor.

What does Resolution and Min.Illumination mean? and how would one interpret with the image quality?


Till now, I was reading the angle(field) of view to be a single dimension quantity. Here they have given it in 2 dimension. Does it mean the horizontal and vertical FOV? If yes, then when I see a single FOV what does it mean? pl. refer http://imaging.nikon.com/lineup/lens...ator/index.htm

If anyone has expertise in thermal imaging, please explain what spectral band is? and how do i correlate it with the image?

Edit:

What does Tri/Quad FOV mean? How can there be more than one FOV?

[Saorsa]

I did some reading and found this link. please refer
http://www.gd-imaging.com/Documents/...DS5-213-10.pdf

Can someone explain the terms in visible camera section. what does 440kpixels in (440k Pixels; 752(H) x 582(V)) mean?

The 440K is rounding of the 752 multiplied by 582. It is the total number of pixels on the sensor. If you have the size of the sensor you will find that you can calculate the pixel size and density which might be useful for comparison.

Till now, I was reading the angle(field) of view to be a single dimension quantity. Here they have given it in 2 dimension. Does it mean the horizontal and vertical FOV? If yes, then when I see a single FOV what does it mean? pl. refer http://imaging.nikon.com/lineup/lens...ator/index.htm

Virtually all lenses create a circular image because it is easier to make circular lenses. Then, we put some sort of light capture device (film or sensor) behind them and make them rectiangular. As a result, they have two dimensions. The longer side will capture a different total angle than the shorter side. Think of it as a cone being projected as a circle onto a drawn rectangle The projected circle that will fit inside the box has a smaller angle than the one that will contain the rectangle.

If anyone has expertise in thermal imaging, please explain what spectral band is? and how do i correlate it with the image?

The spectral band is the range of frequencies that the sensor can capture. A comparison here would be the frequency range of a radio. It represents the colors that you can see in the image. Since our eyes have limitations on what we call visible light, you can only correlate them by processing the image into the range of visible light. That's one reason why thermal images look 'funny'.

[DanK]

The formula would be size of object = tan (angle of view) * distance to object

It's a distant memory for me too, but I think this is not quite right.

A trigonometric function applies to the angles in a right triangle, but this isn't one if the subject is parallel to the sensor. Rather, it's an isosceles triangle. To get a right triangle, you would bisect it in the middle. So I think the formula would be:

size=2tan(a/2)*distance

which is not the same as the formula you gave because as you pointed out, the function isn't linear.

But even that, if it is right, is a simplification. The image collected by the lens is a circle that the sensor truncates to form a rectangle. For that reason, there is not a single angle of view. Rather, the angle is a continuous function bounded at the high end by the angle of view in the corners (where the edge of the rectangle is farthest from the center) and at the low end by the angle of view top to bottom (where the edge of the rectangle is closest). That is why many lens hoods for zoom lenses are petal-shaped. A cylindrical hood that was designed for the narrowest angle of view at the long end would cause horrible vignetting at the short end. But if you design it for the widest angle at the short end, it would be nearly useless, particularly at the long end. So, the petal allows you to get maximum shading while taking the variable angle of view into account.

The new photopills iphone app, which I recently bought but haven't learned to use well, has a field of view/ distance calculator that includes vertical, horizontal, and diagonal angle of view.

[ajohnw]

As far as I know the angle of view of camera lenses is specified across the diagonal of the sensor.

Mentioning reach sounds like an advert type red herring to me. Any lens can be focused on the moon, the stars or anything else. Even infinity according to the markings on some. What is important is magnification - longer focal lengths give more magnification.

Personally I feel that the usual lens spec angle of view doesn't help choose lenses at all. Field glasses have a much better method. The specify the actual width of the field of view at a specific distance. If some one wants to do that on a camera lens they need to ratio the field of view from the diagonal to the dimension across the sensor. Take 1/2 of it and then use tans to calculate 1/2 the actual field of view at what ever distance interests them. On the other hand if they say specified the actual field of view at some convenient distance others could be worked out in the head.

I've always intended to do a spread sheet.

John
-

[DanK]

Mentioning reach sounds like an advert type red herring to me.

Not at all. Even if people call it something else, it is a major factor in choosing lenses, and it can be one in choosing sensor size. That is one reason why the 7D is so popular among birders and wildlife photographers. I think of reach this way: How far can you be from the subject and fill the frame? This is an important consideration. For example, I have two bodies: a new FF, and a 5-year-old crop sensor camera. In most cases, I will choose the newer FF camera. However, when I need maximum reach (for example, when chasing water birds like mergansers in my kayak), I'll take the crop. People will say, 'oh, just crop the image from the FF,' but unless pixel density is similar, that has its costs.

[ajohnw]

Not at all. Even if people call it something else, it is a major factor in choosing lenses, and it can be one in choosing sensor size. That is one reason why the 7D is so popular among birders and wildlife photographers. I think of reach this way: How far can you be from the subject and fill the frame? This is an important consideration. For example, I have two bodies: a new FF, and a 5-year-old crop sensor camera. In most cases, I will choose the newer FF camera. However, when I need maximum reach (for example, when chasing water birds like mergansers in my kayak), I'll take the crop. People will say, 'oh, just crop the image from the FF,' but unless pixel density is similar, that has its costs.

A reach of 9km is meaningless with the magnification being specified. Just about any lens will focus to 9km even though it will be rather close to the infinity focus. If the magnification is specified it still isn't of much direct use. On the other hand if at 9km it had a 2m field of view that would be directly meaningful.

I suppose some one could say one format has more "reach" than another but really that only applies when the same focal length is used on differing formats. Still not much use if for instance you wanted to know how much of the frame a 150mm tall bird would take up at some distance.

John
-

[DanK]

A reach of 9km is meaningless with the magnification being specified. Just about any lens will focus to 9km even though it will be rather close to the infinity focus.

You seem to be confusing focal length with focus. Reach has nothing to do with infinity focus. If you ask "how far away can I be and still fill the frame with a subject of one given size," each focal length will have one solution to that problem for a given sensor size and subject size. If you have a shorter focal length, you have to be closer. If you keep the focal length the same and increase sensor size, you have to be closer. That is what people (at least I) mean by "reach." Reach in this sense is an exact function of field of view, using the equation I gave earlier, so if field of view is meaningful, as you say, then so is reach.

[davidedric]

A trigonometric function applies to the angles in a right triangle, but this isn't one if the subject is parallel to the sensor. Rather, it's an isosceles triangle. To get a right triangle, you would bisect it in the middle. So I think the formula would be:

Correct

[Saorsa]

I'm not sure what the previous parts of this discussion may have entailed. In terms of 'reach' one element is the focal length of the lens and another sensor characteristics.

Some of this points to resolution in terms of a thermal imaging process.

[DanK]

In terms of 'reach' one element is the focal length of the lens and another sensor characteristics.

There is only one characteristic of the sensor that matters: its size. The angle of view is this, where d is whichever dimension of the sensor you are using and F is focal length:

a=2arctan(d/2F)

And the angle of view determines reach. Resolution does not come into play.

For a while, Marc Levoy at Stanford posted on line pdfs from one of his classes that go through these and other related equations. Might still be there.

[Saorsa]

There is only one characteristic of the sensor that matters: its size. The angle of view is this, where d is whichever dimension of the sensor you are using and F is focal length:

a=2arctan(d/2F)

And the angle of view determines reach. Resolution does not come into play.

For a while, Marc Levoy at Stanford posted on line pdfs from one of his classes that go through these and other related equations. Might still be there.

True as far as magnification is concerned. In terms of thermal imaging the sensitivity of the sensor to specific wavelengths needs to be considered or the image is lost in the noise.

[ajohnw]

http://www.oxforddictionaries.com/de.../english/reach

-

[BUGSBUNNYBOSS]

The 440K is rounding of the 752 multiplied by 582. It is the total number of pixels on the sensor. If you have the size of the sensor you will find that you can calculate the pixel size and density which might be useful for comparison.


Virtually all lenses create a circular image because it is easier to make circular lenses. Then, we put some sort of light capture device (film or sensor) behind them and make them rectiangular. As a result, they have two dimensions. The longer side will capture a different total angle than the shorter side. Think of it as a cone being projected as a circle onto a drawn rectangle The projected circle that will fit inside the box has a smaller angle than the one that will contain the rectangle.



The spectral band is the range of frequencies that the sensor can capture. A comparison here would be the frequency range of a radio. It represents the colors that you can see in the image. Since our eyes have limitations on what we call visible light, you can only correlate them by processing the image into the range of visible light. That's one reason why thermal images look 'funny'.

Do you have an idea on the min. illumination?

In this example, the sensor is 1/2 inch CCD. How will I calculate its pixel size and its density?

I Understood your explanation on why there was 2 dimensional angle. But, I still don't understand which angle of view would be practically used when mentioning angle of view (as in the second link I posted; the lens simulator).

Regarding the spectral band, 3-5 µm falls in the infrared region of the spectrum. Does it mean, the camera can only capture Infrared signals. If yes, then how does it show visible light? or can a thermal camera differentiate all of the electromagnetic spectrum?

Please correct me if any of my questions are stupid.

[BUGSBUNNYBOSS]

A reach of 9km is meaningless with the magnification being specified. Just about any lens will focus to 9km even though it will be rather close to the infinity focus. If the magnification is specified it still isn't of much direct use. On the other hand if at 9km it had a 2m field of view that would be directly meaningful.

I don't know if you saw that video or not. In case you haven't, it showed the position of the camera, and starts zooming until at the farthest point, a bus or truck is clearly visible. and the distance between them is 9Km (This is what I meant by reach) and the corresponding angle of view was 0.25 deg.

As far as I know the angle of view of camera lenses is specified across the diagonal of the sensor

If this is the case, then why have they mentioned two angle of view in the brochure (0.22° x 0.17°)?

[ajohnw]

If this is the case, then why have they mentioned two angle of view in the brochure (0.22° x 0.17°)?

For camera lenses in general that is unusual. One angle of view is specified and this appears to be across the diagonal which in many ways isn't of much use without some calculations.

There is a way of using it after a fashion with telephoto lenses where this aspect can be important. We all carry a rough 1/2 degree measure around with us - very rough - the thumb at arms length. People can even calibrate their thumb against a full moon which is close to 1/2 degree across by eye.

Going back to your original question which I assumed some one else had answered - is as sensor of this size comparable with say a full frame or even APS dslr - no there just aren't enough pixels. Relate 0.5mp to your screen or look at it terms of the angle covered by a pixel. That limits the resolution - in this case 0.25 degrees / 752 horizontally. If a lens could be found with a 0.25 degrees angle of view for a DSLR the divisor would be several 1000. Basically the IR stuff is low resolution and intended to show outlines rather than features complicated by the fact that heat is being looked at - you might for instance notice noses because they can be cooler than the face but I doubt if you would see nostrils.

To get around same angle of view on a Nikon DX sensored camera you would need a lens with a focal length of about 3000mm. Paparazzi use things like this to obtain this range of focal lengths.



That one, a 12in Meade scmidt cassegrain has a focal length of about 3000mm. Smaller ones with shorter focal lengths are also available. If some one wanted to go in that direction a Meade ETX-125 might be a more manageable arrangement. 2000mm F15 giving about 2/3 of a degree across the frame of a DX sensor. The image might not fully cover it though and sorting out the length of adapters for correct focus can be a problem or even impossible knowing Meade. There are also smaller schmidt cassegrains.



Also apo refractors but the costs would be horrendous at like this focal length as they are generally around F6 so for 3000mm the diameter of the lens would 3000/6=500mm = more or less impossible.

The thermal imaging cameras use a much smaller sensor. A DX sensor is just short of 24mm across. I/2 that and the focal length for the same angle of view also 1/2's. In practice they are probably a lot smaller than than. The 1/2 ing of the focal length also 1/2's the diameter of the optics for the same F ratio.

John
-

[Saorsa]

Do you have an idea on the min. illumination?

In this example, the sensor is 1/2 inch CCD. How will I calculate its pixel size and its density?

I Understood your explanation on why there was 2 dimensional angle. But, I still don't understand which angle of view would be practically used when mentioning angle of view (as in the second link I posted; the lens simulator).

Regarding the spectral band, 3-5 µm falls in the infrared region of the spectrum. Does it mean, the camera can only capture Infrared signals. If yes, then how does it show visible light? or can a thermal camera differentiate all of the electromagnetic spectrum?

Please correct me if any of my questions are stupid.

None of your questions are stupid but they are the sort that could be answered by looking at spec sheets for the sensor or device. All light of all wavelengths will fall on the sensor. What the sensor captures is a function of it's design and that of the device. For example, unwanted frequencies may be optically filtered out so that only the desired range (IR in this case) arrive at the sensor. Further selection and rejection can be done with Digital Signal Processing programs after capture.

As a result, the answers that anyone here can give you are general statements and you seem to be looking for specific points to make a decision or complete a project. There is some ambiguity that makes this a poor method of dealing with questions. Take angle of view. There is one inherent in the lens by it's design. As I mentioned, it is a single number defined by the angle of the cone of light. But, as soon a sensor is used there are angles defined within that based on the height and width of the sensor. Take a look at this comparison of camera sensor sizes.



The lens would have an angle of view that would define a circle that would encompass all of those rectangles. Each rectangle could be considered as having three other angles of view. Horizontal, Vertical and Diagonal based on their ability to capture the light.

Book a conference room with a whiteboard and some dry-erase markers and we could probably knock this out in half an hour.

[ajohnw]

There is another rule of thumb worth remembering. When say as on the Nikon 1 range they mention 1in sensors it's all based on old vidicon tube cameras "sort of like valves". The diagonal size of all of the 1/2 in or what ever will be around 2/3 of that. Format shape varies this but it is very close to 2/3 on traditional formats.

John
-

[Saorsa]

Ooops, I forgot the question on illumination.

Look at the sensor size comparison again and consider that if the lens is not designed to cover the sensor but has a larger angle of view than necessary then some of the light is falling off the edges. That is lost illumination. The right lens will put as much light as possible directly on the sensor.

There has always been discussion of what is a normal lens and the general definition seems to be a lens which accepts an angle of view similar to the human eye and projects it onto the film or sensor. For each type of camera this 'normal' focal length is different. For 35mm 50mm became the normal, for 6x6cm it was 80mm for 4x5 inch it was 135mm, etc. Today, we amplify the confusion with discussion of apparent FL on crop sensor bodies.