Filtering Data which not function of time

Hello, I have a thermal data from infra-red camera.

Uh... Let's suppose this infra red camera has 1*1000 pixels.

With this camera, I have taken a single photo of warm sausage's temperature.

However, this thousand pixels has some noises, so somehow I want to erase the noise.

But this data is not function of time, if they are, It would be easy to deal with noises with many methods like sampling frequency, FTT, low pass... etc.

However this not-function of time-data.... How should I deal with noise these kind of data?

Answer 1

[Star Strider]

The data being filtered do not have to be functions of time. They can be functions of anything, so long as the sampling interval of the independent variable is constant. The frequency of the Fourier transformed data will then by cycles/unit, whatever that unit happens to be. If it is seconds, then cycles/second, or Hz, however if it millimetres, the frequency will be cycles/mm.

The Fourier transform (fft or fft2) cares not at all about the units, only about the data.

That aside, use the imfilter (or filter2) function to do the filtering, and perhaps the fspecial or similar functions to design the filter, if a more sophisticated filter is necesary.

.

[도연 원]

Thanks! I was late.

[Star Strider]

As always, my pleasure!

No worries! Thank you!