Various signal measurements requires to detect and parametrize event in it. Mostly the threshold detection is widely used, which is sufficient method to event detection. But for better event parametrization the more precise event start within signal sometimes is needed. Thus, the Retegate.SignalEventDetection has been developed.
- Detect raw events in signal via threshold detection
- Use the detected events as input for the SignalEventDetection
- Get the precise event start within the detected events
Note: When none or empty signal provided, the ArgumentNullException is thrown. Signal with the range <0,1> is expected, otherwise ArgumentException is thrown.
Please, be advised that the event start is detected from the signal part starting with 0 sample index and ending in maximum signal peak. I.e. for constant and declining signal the detection fallback to 0 index sample. (If you need similar way to detect more precisely the end of event, you need to transform the signal - reverse it and use the SignalEventDetection again, and reverse the result also vice-versa.)
using Retegate.SignalEventDetection;
var eventStartSample = SignalEventDetection.GetIndexOfEventStart(inputData);The mathematical principle behind the SignalEventDetection is based on the Akaike Information Criterion (AIC). It is widely used in Acoustic Emission (AE) in Non-Destructive Testing (NDT) and also in seismology. (Personally I use it also for produced part detection within continuous image for visual quality check and for better automated histogram range choosing)
Feel free to use this code in your projects (it is under MIT licence). If you have any questions, please contact me at retegate@retegate.com