Want to remove 'noise' from a matrix.

I have a single column matrix with values ranging from around -15 to 50, there's around 19000 values and I wish to remove any values below 20 and set them as 0. This is what I currently have to import the data:

input = xlsread('HiG_67.csv','J:J');

Thanks

[Stephen23]

Do not call your variable input as this is the name of a commonly used inbuilt function input. When you shadow the name of an inbuilt function like this it stops the inbuilt function from working.

For this reason never use the names size, length, input, i, j, cat, length, etc. You can use which to check if a name is already used-

Answer 1

[Dennie]

If input is a non-negative single column then you can use the following simple filter:

FilterValue=20;

input(input<FilterValue)=0;

If there are also negative values that you wish to keep then you can try:

input(abs(input)<FilterValue)=0;

If you only want to filter the positive noise:

filter=input<FilterValue & input>=0;
input(filter))=0;

Dennie

[OliK]

If now I have say 30 'peaks' in my data how would I create a new matrix with only peaks 10-19 .

[Image Analyst]

First find the peaks. findpeaks() in the Signal Processing Toolbox will help. Then, once you know the peaks go from, say, index 42 to index 173, just extract those indices:

peaks10To19 = allPeaks(index1 : index2);

where allPeaks is the name of your data array (like Thorsten said, it better not be "input"!!!).

[OliK]

Is there a way I can do it without Signal Processing Toolbox?

[Image Analyst]

There are peak detectors in the File Exchange.

Answer 2

[Thorsten]

 x(x < 20) = 0;

BTW: Please don't call your variable input, it's a Matlab function.