DSP problem in Labview

[mugger]

I have a DSP problem to be relized in the labview. However, I'm not quite familiar with some toolboxes. The goal is to detact the pulse including the length and intensity from a real-time noisy signal. Given the pulse signal, how to locate the breakdown points? 

Thanks

[johnsold · Accepted solution]

It appears that the signal at the large transitions is larger than most of the noise.  I think I might try looking for large transitions, say >8-10 in 2-3 samples.  When I find such a jump, then look at the mean before and after to see if is a real pulse or just noise.

If the pulse widths are known or have a narrow range of values, you can also use that information to help discriminate between signal and noise.  The very large transitions bounding the negative part of the pulse are much easier to detect.

Lynn

[mugger]

Do you know any built-in subVIs that could auto-detect these transitions? or jumps?

[johnsold]

No.  I think you will need to write your own.  It should not be too difficult to get started.  It may take some fine tuning later, depending on how well-behaved the data sets are.

Lynn