calculate signal power of image

16 次查看(过去 30 天)
how to calculate signal power of image in matlab?
  4 个评论
Walter Roberson
Walter Roberson 2013-1-5
Okay, give us the particular equation you need to use and we can tell you something about how to implement it.
But if what you are saying is that you are looking for the equation for "signal power of image", our answer is "There are too many things that sometimes get called that, and you will need to be more specific about which measure of signal power you need." There is no "default" or "nearly standard" meaning for "signal power for image". There are more specific terms such as SNR and PSNR
vipul utsav
vipul utsav 2013-1-5
for SNR calculations i.e SNR=P(signal)/P(noise)

请先登录,再进行评论。

采纳的回答

Image Analyst
Image Analyst 2013-1-5
vipul, there is no power in an image, by the strict definition, but there is energy. You may have some other layman's definition. If so, let's hear it. Let's look at the units. I'll use SI units here. Power is in watts, or joules per second. A light source has power and it emits it, and it lands on your scene which then reflects it onto your camera sensor. So what hits a given pixel? It's called irradiance and its units are power per unit area or watts/m^2 or joules/(seconds*m^2). By the way, you can look up any of these radiometric definitions in Wikipedia or here
So what does a gray level really mean? The gray level is a measure of energy? Why energy? Exposure is the irradiance integrated over time and has units of joules per square meter (the "per second" units got integrated away). It's the measure of the energy of all the photons that hit the area of sensor. When you look at exposure for one single pixel, then you sum up (integrate) the exposure over the area of that one pixel, so the meters squared go away and you are left with joules. So gray level is proportional to joules, or energy. It's also proportional to power since it's power integrated over the area of a pixel and over the time of exposure, but both of those are constants for a given snapshot. So the pixel value is proportional to power by a conversion factor (that has units), but it really has units of energy.
So when you see SNR in the Wikipedia article, the gray level goes in place of the P (Power) in the equation, not in place of the A (amplitude). And when you see formulas with 10Log or 20Log, you use the 10Log formula, not the 20Log formula.
I hope that explains it better.
  2 个评论
vipul utsav
vipul utsav 2013-1-6
编辑:vipul utsav 2013-1-6
you mean,pixel intensity=radiance=W/cm^2=power/area
Image Analyst
Image Analyst 2013-1-6
编辑:Image Analyst 2013-1-6
No! Radiance is an entirely different animal. Radiance is a measure of how much light a source emits, not of how much light hits a scene. The SI unit of radiance is watts per steradian per square meter (W·sr−1·m−2). I know the units used in optics and radiometry can be very confusing, so you have to read carefully. Irradiance is W/m^2 but a pixel has an area and an integration time (like 1/30th of a second), so W/m^2 = Joules/(second*m^2) but you integrate over time and area so you're left with joules, a unit of energy, not power and not irradiance and certainly not radiance. Radiance and irradiance are complementary concepts in general (any part of the spectrum). If you are using just the visible part of the spectrum and weight it by the luminous efficacy of the human eye, then you're talking about luminance and illuminance. Note that both irradiance and illuminance start with "i" which means that they refer to quantities that are "incident" on a surface.

请先登录,再进行评论。

更多回答(1 个)

Walter Roberson
Walter Roberson 2013-1-5

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by