selecting and extracting contiguous points as one event

Question

Stephen 2019-1-7

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/438683-selecting-and-extracting-contiguous-points-as-one-event

评论： Stephen 2019-1-8

sample responses.xlsx

I'm sure there is probably a drop-dead simple way to do this, but my brain is fried at the moment and I can't think of one. I'm trying to identify and count numbers of events that exceed a certain value in a time series. For example, a plot of the attached file yields:

I'd like to be able to count the number of events that exceed the red line (y=0.13). Sometimes there is 1 point > 0.13 (e.g. the first event above at x=40) and that is easy to count as a single signal. But sometimes there are 2 or more contiguous points > 0.13 (e.g. the second event above, at x=56, 57), yet should be counted as just one "event", not as 2 events.

How can this be done? I've attached a file with the above signal.

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

nanren888 2019-1-7

Perhaps you could count the number of upward level-crossings?

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Andrei Bobrov 2019-1-7

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/438683-selecting-and-extracting-contiguous-points-as-one-event#answer_355409

在 MATLAB Online 中打开

let a - your data - vector.

out = sum(diff([0;a(:) > .13]) == 1);

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Stephen 2019-1-7

Oh wow, GREAT! Thanks Andrei!

请先登录，再进行评论。

Answer 2

nanren888 2019-1-7

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/438683-selecting-and-extracting-contiguous-points-as-one-event#answer_355407

tt = linspace(0,12,1001);

s= sin(tt);

th = 0.4;

nnz(diff(s-th));

dd = [0,diff((s-th)>0)>0.5];

plot(tt,s,tt,dd);

grid('on');

nnz(dd)

ans =

2

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Stephen 2019-1-7

Thanks, nanren888! I'm having trouble applying your solution to a matrix of numbers (such as the excel spreadsheet that was attached with my Q) instead of a mathematical function such as the sin wave you show. sorry, I'm not as advanced as you are.

请先登录，再进行评论。