On scipy.signal.find_peaks ()
The official document is still ahead scipy.signal.find_peaks
Due to the need to monitor the peak value of the waveform, so finding the function
The function peak is found by comparing the position of the surrounding
Input:
x: peak signal sequence with
height: signal falls below a specified height is not considered
threshold: the vertical distance between neighboring samples
distance: minimum horizontal distance between adjacent peaks, first remove the smaller peak until all conditions of the remaining peaks are satisfied.
prominence: the degree of personal understanding is a projection, see peak_prominences
width: the width of the peak, see peak_widths
plateau_size : the number of guaranteed flattened peak is greater than a given value corresponding to
Output:
Peaks : peak index corresponding to x
properties:
height--> ‘peak_heights’
threshold-->‘left_thresholds’, ‘right_thresholds’
prominence-->‘prominences’, ‘right_bases’, ‘left_bases’
width-->‘width_heights’, ‘left_ips’, ‘right_ips’
plateau_size-->‘plateau_sizes’, left_edges’, ‘right_edges’
For the case of a large noise, smoothing the signal should then take the peak, or a wavelet transform find_peaks_cwt
to achieve peak lookup