I recorded the raw response signal and repeated the set of experiments to reduce uncertainty in the measurement by averaging process.
I always use Gaussian Smoothing Filter. You can find more information about this filtering procedure in my article.
Not sure what you have as data, and also do you have one set of independent data points or two measures of the same data?
If you have a 1D (time meaure) or 2D (image) signal of some data that has some local correlation, meaning you can model how smooth it should be them some sort of smoothing can be used. In that case Gaussian, Weiner or moving average filtering can be used. Look into anisotropic filtering for a more robust edge preservation in images.
If you have several measurements of the same data (like multiple images of the same object) and want to remove the noise then average of median filtering over the multiple samples is the easiest choice.
As a more advanced choice look into non-local means filtering (if you don't care about non-linearity in your processing pipeline):
http://www.mathworks.com/matlabcentral/fileexchange/13176-non-local-means-filter
For images use the 3D FFT version of non-local mean filtering:
http://www.cs.tut.fi/~foi/GCF-BM3D/
The answer to your question depends on the application and context in which you have acquired the data. So in order to get the relevant response, please provide some details regarding the purpose and application for which you are collecting the data.
Actually I measured acceleration response from an accelerometer and impulse response response from hammer through DAQ. I wish to find out mode shape from modal analysis experiment.
So you are doing Impact modal test :)
The standard way is to average in frequency domain. So after each measurement, convert your raw time data to frequency domain by using FFT, calculate auto and cross spectrum and average them. Typically, 5 averages should be sufficient (in other words, you need to repeat each measurement 5 times).
I'd do as Shashank Chauhan suggested.I think this is the correct procedure.
- Badges
- Science topic
- Similar topics
- Engineering
- Electrical Engineering
- Signal Processing