Data not being filtered out by threshold

Question

Joshua Isherwood 2017-5-10

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/339566-data-not-being-filtered-out-by-threshold

关闭： MATLAB Answer Bot 2021-8-20

pressure.mat

I am trying to create touch down and take off events on my centre of pressure data, however the threshold I have created seems to consistently come in one frame too early.

The threshold is allowing data below (-0.51) the set value (<2) through. Even when the threshold is increased this is happening. So i am constantly left with a negative value as first touch down before the actual touch down occurs one frame later.

This is the code i am using to create the events:

Air = zeros(length(mm3),1);
for n = 1:length(mm3)
    if mm3(n,2) < 2
        Air(n,1) = 1;
    end
end
Edges = diff(Air);             % Steigung kennzeichnet aufsteigende und abfallende Kante
TOPressure = find(Edges==1);     % Absprungzeitpunkte
TDPressure = find(Edges==-1);
plot(mm3(:,2))
TOPressure = TOPressure(5:55,1);
TDPressure = TDPressure(5:55,1);
TOPressure = TOPressure(2:47,1);
TDPressure = TDPressure(1:46,1);

2 个评论
显示无隐藏无

Jan 2017-5-10

在 MATLAB Online 中打开

I do not see a relation between the description and the code. The code does not contain any threshold. What does "seems to come one frame too early" exactly mean? Which variable is concerned?

By the way: You can create Air directly without pre-allocation and a loop:

Air = double(mm3(:, 2) < 2);

Joshua Isherwood 2017-5-11

The threshold would be this line : if mm3(n,2) < 2 so the code should create the events TDPressure and TDPressure at the first values exceeding 2.

Sorry a little vague on that part. The variable is the centre of pressure measured during running so the events are supposed to be created during initial ground contact (TDPressure) and final ground contact (TOPressure) however the events are being created one frame too early with the foot still in the air.

So it allows one frame of data < 2 through and creates the events then as apposed to the following frame which is actually > 2.

Have included two files, mm3 which shows that the initial contact occurs at frame 206 with a value > 2, then the TDPressure which has the initial contact occurring at 205 with a value at < 2.

此问题已关闭。

Answer 1

Santhana Raj 2017-5-11

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/339566-data-not-being-filtered-out-by-threshold#answer_266551

If you see the help of diff, it says: diff(X), for a vector X, is [X(2)-X(1) X(3)-X(2) ... X(n)-X(n-1)].

So, in your case, 80th element of Air is actually 1. And when you use diff, 79th element of TOPressure is Air(80)-Air(79), (which is actually the edge).