Numpy Correlate is not providing an offset - python

I am trying to look at astronomical spectra using Python, and I'm using numpy.correlate to try and find a radial velocity shift. I'm comparing each spectrum I have to one template spectrum. The problem that I'm encountering is that, no matter which spectra I use, numpy.correlate states that the maximal value of the correlation function occurs with a shift of zero pixels, i.e. the spectra already line up, which is very clearly not true. Here is some of the relevant code:
corr = np.correlate(temp_data, imag_data, mode='same')
ax1.plot(delta_data, corr, c='g')
ax1.plot(delta_data, 100*temp_data, c='b')
ax1.plot(delta_data, 100*imag_data, c='r')
The output of this code is shown here:
What I Have
Note that the cross-correlation function peaks at an offset of zero pixels despite the template (blue) and observed (red) spectra clearly showing an offset. What I would expect to see would be something a bit like (albeit not exactly like; this is merely the closest representation I could produce):
What I Want
Here I have introduced an artificial offset of 50 pixels in the template data, and they more or less line up now. What I would like is, for a case like this, for a peak to appear at an offset of 50 pixels rather than at zero (I don't care if the spectra at the bottom appear lined up; that is merely for visual representation). However, despite several hours of work and research online, I can't find someone who even describes this problem, let alone a solution. I've attempted to use ScyPy's correlate and MatLib's xcorr, and bot show this same thing (although I'm led to believe that they are essentially the same function).
Why is the cross-correlation not acting the way I expect, and how to do I get it to act in a useful way?

The issue you're experiencing is probably because your spectra are not zero-centered; their RMS value looks to be about 100 in whichever units you're plotting, instead of 0. The reason this is an issue is because numpy.correlate works by "sliding" imag_data over temp_data to get their dot product at each possible offset between the two series. (See the wikipedia on cross-correlation to understand the operation itself.) When using mode='same' to produce an output that is the same length as your first input (temp_data), NumPy has to "pad" a bunch of dummy values--zeroes--to the ends of imag_data in order to be able to calculate the dot products of all the shifted versions of the imag_data. When we have any non-zero offset between the spectra, some of the values in temp_data are being multiplied by those dummy zero-padding values instead of the values in image_data. If the values in the spectra were centered around zero (RMS=0), then this zero-padding would not impact our expectation of the dot product, but because these spectra have RMS values around 100 units, that dot product (our correlation) is largest when we lay the two spectra on top of one another with no offset.
Notice that your cross-correlation result looks like a triangular pulse, which is what you might expect from the cross-correlation of two square pulses (c.f. Convolution of a Rectangular "Pulse" With Itself. That's because your spectra, once padded, look like a step function from zero up to a pulse of slightly noisy values around 100. You can try convolving with mode='full' to see the entire response of the two spectra you're correlating, or, notice that with mode='valid' that you should only get one value in return, since your two spectra are the exact same length, so there is only one offset (zero!) where you can entirely line them up.
To sidestep this issue, you can try either subtracting away the RMS value of the spectra so that they are zero-centered, or manually padding the beginning and end of imag_data with (len(temp_data)/2-1) dummy values equal to np.sqrt(np.mean(imag_data**2))
Edit:
In response to your questions in the comments, I thought I'd include a graphic to make the point I'm trying to describe a little clearer.
Say we have two vectors of values, not entirely unlike your spectra, each with some large non-zero mean.
# Generate two noisy, but correlated series
t = np.linspace(0,250,250) # time domain from 0 to 250 steps
# signal_model = narrow_peak + gaussian_noise + constant
f = 10*np.exp(-((t-90)**2)/8) + np.random.randn(250) + 40
g = 10*np.exp(-((t-180)**2)/8) + np.random.randn(250) + 40
f has a spike around t=90, and g has a spike around t=180. So we expect the correlation of g and f to have a spike around a lag of 90 timesteps (in the case of spectra, frequency bins instead of timesteps.)
But in order to get an output that is the same shape as our inputs, as in np.correlate(g,f,mode='same'), we have to "pad" f on either side with half its length in dummy values: np.correlate pads with zeroes. If we don't pad f (as in np.correlate(g,f,mode='valid')), we will only get one value in return (the correlation with zero offset), because f and g are the same length, and there is no room to shift one of the signals relative to the other.
When you calculate the correlation of g and f after that padding, you find that it peaks when the non-zero portion of signals aligns completely, that is, when there is no offset between the original f and g. This is because the RMS value of the signals is so much higher than zero--the size of the overlap of f and g depends much more strongly on the number of elements overlapping at this high RMS level than on the relatively small fluctuations each function has around it. We can remove this large contribution to the correlation by subtracting the RMS level from each series. In the graph below, the gray line on the right shows the cross-correlation the two series before zero-centering, and the teal line shows the cross-correlation after. The gray line is, like your first attempt, triangular with the overlap of the two non-zero signals. The teal line better reflects the correlation between the fluctuation of the two signals, as we desired.
xcorr = np.correlate(g,f,'same')
xcorr_rms = np.correlate(g-40,f-40,'same')
fig, axes = plt.subplots(5,2,figsize=(18,18),gridspec_kw={'width_ratios':[5,2]})
for n, axis in enumerate(axes):
offset = (0,75,125,215,250)[n]
fp = np.pad(f,[offset,250-offset],mode='constant',constant_values=0.)
gp = np.pad(g,[125,125],mode='constant',constant_values=0.)
axis[0].plot(fp,color='purple',lw=1.65)
axis[0].plot(gp,color='orange',lw=lw)
axis[0].axvspan(max(125,offset),min(375,offset+250),color='blue',alpha=0.06)
axis[0].axvspan(0,max(125,offset),color='brown',alpha=0.03)
axis[0].axvspan(min(375,offset+250),500,color='brown',alpha=0.03)
if n==0:
axis[0].legend(['f','g'])
axis[0].set_title('offset={}'.format(offset-125))
axis[1].plot(xcorr/(40*40),color='gray')
axis[1].plot(xcorr_rms,color='teal')
axis[1].axvline(offset,-100,350,color='maroon',lw=5,alpha=0.5)
if n == 0:
axis[1].legend(["$g \star f$","$g' \star f'$","offset"],loc='upper left')
plt.show()

Related

Fourier Transform Time Series in Python

I've got a time series of sunspot numbers, where the mean number of sunspots is counted per month, and I'm trying to use a Fourier Transform to convert from the time domain to the frequency domain. The data used is from https://wwwbis.sidc.be/silso/infosnmtot.
The first thing I'm confused about is how to express the sampling frequency as once per month. Do I need to convert it to seconds, eg. 1/(seconds in 30 days)? Here's what I've got so far:
fs = 1/2592000
#the sampling frequency is 1/(seconds in a month)
fourier = np.fft.fft(sn_value)
#sn_value is the mean number of sunspots measured each month
freqs = np.fft.fftfreq(sn_value.size,d=fs)
power_spectrum = np.abs(fourier)
plt.plot(freqs,power_spectrum)
plt.xlim(0,max(freqs))
plt.title("Power Spectral Density of the Sunspot Number Time Series")
plt.grid(True)
I don't think this is correct - namely because I don't know what the scale of the x-axis is. However I do know that there should be a peak at (11years)^-1.
The second thing I'm wondering from this graph is why there seems to be two lines - one being a horizontal line just above y=0. It's more clear when I change the x-axis bounds to: plt.xlim(0,1).
Am I using the fourier transform functions incorrectly?
You can use any units you want. Feel free to express your sampling frequency as fs=12 (samples/year), the x-axis will then be 1/year units. Or use fs=1 (sample/month), the units will then be 1/month.
The extra line you spotted comes from the way you plot your data. Look at the output of the np.fft.fftfreq call. The first half of that array contains positive values from 0 to 1.2e6 or so, the other half contain negative values from -1.2e6 to almost 0. By plotting all your data, you get a data line from 0 to the right, then a straight line from the rightmost point to the leftmost point, then the rest of the data line back to zero. Your xlim call makes it so you don’t see half the data plotted.
Typically you’d plot only the first half of your data, just crop the freqs and power_spectrum arrays.

Understanding the output of fftfreq function and the fft plot for a single row in an image

I am trying to understand the function fftfreq and the resulting plot generated by adding real and imaginary components for one row in the image. Here is what I did:
import numpy as np
import cv2
import matplotlib.pyplot as plt
image = cv2.imread("images/construction_150_200_background.png", 0)
image_fft = np.fft.fft(image)
real = image_fft.real
imag = image_fft.imag
real_row_bw = image_fft[np.ceil(image.shape[0]/2).astype(np.int),0:image.shape[1]]
imag_row_bw = image_fft[np.ceil(image.shape[0]/2).astype(np.int),0:image.shape[1]]
sum = real_row_bw + imag_row_bw
plt.plot(np.fft.fftfreq(image.shape[1]), sum)
plt.show()
Here is image of the plot generated :
I read the image from the disk, calculate the Fourier transform and extract the real and imaginary parts. Then I sum the sine and cosine components and plot using the pyplot library.
Could someone please help me understand the fftfreq function? Also what does the peak represent in the plot for the following image:
I understand that Fourier transform maps the image from spatial domain to the frequency domain but I cannot make much sense from the graph.
Note: I am unable to upload the images directly here, as at the moment of asking the question, I am getting an upload error.
I don't think that you really need fftfreq to look for frequency-domain information in images, but I'll try to explain it anyway.
fftfreq is used to calculate the frequencies that correspond to each bin in an FFT that you calculate. You are using fftfreq to define the x coordinates on your graph.
fftfreq has two arguments: one mandatory, one optional. The mandatory first argument is an integer, the window length you used to calculate an FFT. You will have the same number of frequency bins in the FFT as you had samples in the window. The optional second argument is the time period per window. If you don't specify it, the default is a period of 1. I don't know whether a sample rate is a meaningful quantity for an image, so I can understand you not specifying it. Maybe you want to give the period in pixels? It's up to you.
Your FFT's frequency bins start at the negative Nyquist frequency, which is half the sample rate (default = -0.5), or a little higher; and it ends at the positive Nyquist frequency (+0.5), or a little lower.
The fftfreq function returns the frequencies in a funny order though. The zero frequency is always the zeroth element. The frequencies count up to the maximum positive frequency, and then flip to the maximum negative frequency and count upwards towards zero. The reason for this strange ordering is that if you're doing FFT's with real-valued data (you are, image pixels do not have complex values), the negative frequency data is exactly equal to the corresponding positive frequency data and is redundant. This ordering makes it easy to throw the negative frequencies away: just take the first half of the array. Since you aren't doing that, you're plotting the negative frequencies too. If you should choose to ignore the second half of the array, the negative frequencies will be removed.
As for the strong spike that you see at the zero frequency in your image, this is probably because your image data is RGB values which range from 0 to 255. There's a huge "DC offset" in your data. It looks like you're using Matplotlib. If you are plotting in an interactive window, you can use the zoom rectangle to look at that horizontal line. If you push the DC offset off scale, setting the Y axis scale to perhaps ±500, I bet you will start to see that the horizontal line isn't exactly horizontal after all.
Once you know which bin contains your DC offset, if you don't want to see it, you can just assign the value of the fft in that bin to zero. Then the graph will scale automatically.
By the way, these two lines of code perform identical calculations, so you aren't actually taking the sine and cosine components like your text says:
real_row_bw = image_fft[np.ceil(image.shape[0]/2).astype(np.int),0:image.shape[1]]
imag_row_bw = image_fft[np.ceil(image.shape[0]/2).astype(np.int),0:image.shape[1]]
And one last thing: to sum the sine and cosine components properly (once you have them), since they're at right angles, you need to use a vector sum rather than a scalar sum. Look at the function numpy.linalg.norm.

trouble with scipy interpolation

I'm having trouble using the scipy interpolation methods to generate a nice smooth curve from the data points given. I've tried using the standard 1D interpolation, the Rbf interpolation with all options (cubic, gaussian, multiquadric etc.)
in the image provided, the blue line is the original data, and I'm looking to first smooth the sharp edges, and then have dynamically editable points from which to recalculate the curve. Each time a single point is edited it should auto calculate a new spline of some sort to smoothly transition between each point.
It kind of works when the points are within a particular range of each other as below.
But if the points end up too far apart, or too close together, I end up with issues like the following.
Key points are:
The curve MUST be flat between the first two points
The curve must NOT go below point 1 or 2 (i.e. derivative can't be negative)
~15 points (not shown) between points 2 and 3 are also editable and the line between is not necessarily linear. Full control over each of these points is a must, as is the curve going through each of them.
I'm happy to break it down into smaller curves that i then join/convolve, but just need to ensure a >0 gradient.
sample data:
x=[0, 37, 50, 105, 115,120]
y=[0.00965, 0.00965, 0.047850827205882, 0.35600416666667, 0.38074375, 0.38074375]
As an example, try moving point 2 (x=37) to an extreme value, say 10 (keep y the same). Just ensure that all points from x=0 to x=10 (or any other variation) have identical y values of 0.00965.
any assistance is greatly appreciated.
UPDATE
Attempted pchip method suggested in comments with the results below:
pchip method, better and worse...
Solved!
While I'm not sure that this is exactly true, it is as if the spline tools for creating Bezier curves treat the control points as points the calculated curve must go through - which is not true in my case. I couldn't figure out how to turn this feature off, so I found the cubic formula for a Bezier curve (cubic is what I need) and calculated my own points. I only then had to do a little adjustment to make the points fit the required integer x values - in my case, near enough is good enough. I would otherwise have needed to interpolate linearly between two points either side of the desired x value and determine the exact value.
For those interested, cubic needs 4 points - start, end, and 2 control points. The rule is:
B(t) = (1-t)^3 P0 + 3(1-t)^2 tP1 + 3(1-t)t^2 P2 + t^3 P3
Calculate for x and y separately, using a list of values for t. If you need to gradient match, just make sure that the control points for P1 and P2 are only moved along the same gradient as the preceding/proceeding sections.
Perfect result

Fitting shifted and deformed dataset to dataset in python

I am trying to analyze the shift and deforming of a pressure test on a block of concrete.
I have two measurements: one of a height (vector h) and one of the diameter (vector d) at that specific height. I have these measurements for different pressures deforming the block. I will denote the height and diameter for different pressures with the index i (h_1 to h_n, d_i respectively). len(h_i) is not necessarily equal to len(h_j)
I want to find a way to fit (deform and shift) the graph
G_1 = (h_1, d_1) to the graph
G_2 = (h_2, d_2)
I thought of minimizing the square error like it is done in functional fitting, I do however have some problems:
I do not know how to introduce the shift in the height/ x-direction
i.e. f(h_(i+1))=f(h_i - h_shift)
I do not know how to introduce the compression in the height/ x-direction
i.e. f(h_(i+1))=f(a*h_i )
I am not sure how I can introduce the "normal" fitting deforming diameter depending on the height. (say I want to add a deforming of the form d_(i+1)=d_i+ah²+bh+c)
I want to stress again that I do not try to manipulate a function to fit data-points but that I try to manipulate a set of points to fit a different set of points.
UPDATE: I have stored illustrations here on Google Drive
Note that there are two different kinds of shift: one small one (sample_fitting_1.png) and one large one (sample_fitting_2.png).
I am trying not to loose the fine structure of the data as I would by fitting a curve through it.
The goal is to shift and deform one of the graphs onto the other one by manipulating it as described above as well in x as in y-direction.
Thanks in advance
Stephan

How to correlate two time series with gaps and different time bases?

I have two time series of 3D accelerometer data that have different time bases (clocks started at different times, with some very slight creep during the sampling time), as well as containing many gaps of different size (due to delays associated with writing to separate flash devices).
The accelerometers I'm using are the inexpensive GCDC X250-2. I'm running the accelerometers at their highest gain, so the data has a significant noise floor.
The time series each have about 2 million data points (over an hour at 512 samples/sec), and contain about 500 events of interest, where a typical event spans 100-150 samples (200-300 ms each). Many of these events are affected by data outages during flash writes.
So, the data isn't pristine, and isn't even very pretty. But my eyeball inspection shows it clearly contains the information I'm interested in. (I can post plots, if needed.)
The accelerometers are in similar environments but are only moderately coupled, meaning that I can tell by eye which events match from each accelerometer, but I have been unsuccessful so far doing so in software. Due to physical limitations, the devices are also mounted in different orientations, where the axes don't match, but they are as close to orthogonal as I could make them. So, for example, for 3-axis accelerometers A & B, +Ax maps to -By (up-down), +Az maps to -Bx (left-right), and +Ay maps to -Bz (front-back).
My initial goal is to correlate shock events on the vertical axis, though I would eventually like to a) automatically discover the axis mapping, b) correlate activity on the mapped aces, and c) extract behavior differences between the two accelerometers (such as twisting or flexing).
The nature of the times series data makes Python's numpy.correlate() unusable. I've also looked at R's Zoo package, but have made no headway with it. I've looked to different fields of signal analysis for help, but I've made no progress.
Anyone have any clues for what I can do, or approaches I should research?
Update 28 Feb 2011: Added some plots here showing examples of the data.
My interpretation of your question: Given two very long, noisy time series, find a shift of one that matches large 'bumps' in one signal to large bumps in the other signal.
My suggestion: interpolate the data so it's uniformly spaced, rectify and smooth the data (assuming the phase of the fast oscillations is uninteresting), and do a one-point-at-a-time cross correlation (assuming a small shift will line up the data).
import numpy
from scipy.ndimage import gaussian_filter
"""
sig1 and sig 2 are assumed to be large, 1D numpy arrays
sig1 is sampled at times t1, sig2 is sampled at times t2
t_start, t_end, is your desired sampling interval
t_len is your desired number of measurements
"""
t = numpy.linspace(t_start, t_end, t_len)
sig1 = numpy.interp(t, t1, sig1)
sig2 = numpy.interp(t, t2, sig2)
#Now sig1 and sig2 are sampled at the same points.
"""
Rectify and smooth, so 'peaks' will stand out.
This makes big assumptions about your data;
these assumptions seem true-ish based on your plots.
"""
sigma = 10 #Tune this parameter to get the right smoothing
sig1, sig2 = abs(sig1), abs(sig2)
sig1, sig2 = gaussian_filter(sig1, sigma), gaussian_filter(sig2, sigma)
"""
Now sig1 and sig2 should look smoothly varying, with humps at each 'event'.
Hopefully we can search a small range of shifts to find the maximum of the
cross-correlation. This assumes your data are *nearly* lined up already.
"""
max_xc = 0
best_shift = 0
for shift in range(-10, 10): #Tune this search range
xc = (numpy.roll(sig1, shift) * sig2).sum()
if xc > max_xc:
max_xc = xc
best_shift = shift
print 'Best shift:', best_shift
"""
If best_shift is at the edges of your search range,
you should expand the search range.
"""
If the data contains gaps of unknown sizes that are different in each time series, then I would give up on trying to correlate entire sequences, and instead try cross correlating pairs of short windows on each time series, say overlapping windows twice the length of a typical event (300 samples long). Find potential high cross correlation matches across all possibilities, and then impose a sequential ordering constraint on the potential matches to get sequences of matched windows.
From there you have smaller problems that are easier to analyze.
This isn't a technical answer, but it might help you come up with one:
Convert the plot to an image, and stick it into a decent image program like gimp or photoshop
break the plots into discrete images whenever there's a gap
put the first series of plots in a horizontal line
put the second series in a horizontal line right underneath it
visually identify the first correlated event
if the two events are not lined up vertically:
select whichever instance is further to the left and everything to the right of it on that row
drag those things to the right until they line up
This is pretty much how an audio editor works, so you if you converted it into a simple audio format like an uncompressed WAV file, you could manipulate it directly in something like Audacity. (It'll sound horrible, of course, but you'll be able to move the data plots around pretty easily.)
Actually, audacity has a scripting language called nyquist, too, so if you don't need the program to detect the correlations (or you're at least willing to defer that step for the time being) you could probably use some combination of audacity's markers and nyquist to automate the alignment and export the clean data in your format of choice once you tag the correlation points.
My guess is, you'll have to manually build an offset table that aligns the "matches" between the series. Below is an example of a way to get those matches. The idea is to shift the data left-right until it lines up and then adjust the scale until it "matches". Give it a try.
library(rpanel)
#Generate the x1 and x2 data
n1 <- rnorm(500)
n2 <- rnorm(200)
x1 <- c(n1, rep(0,100), n2, rep(0,150))
x2 <- c(rep(0,50), 2*n1, rep(0,150), 3*n2, rep(0,50))
#Build the panel function that will draw/update the graph
lvm.draw <- function(panel) {
plot(x=(1:length(panel$dat3))+panel$off, y=panel$dat3, ylim=panel$dat1, xlab="", ylab="y", main=paste("Alignment Graph Offset = ", panel$off, " Scale = ", panel$sca, sep=""), typ="l")
lines(x=1:length(panel$dat3), y=panel$sca*panel$dat4, col="red")
grid()
panel
}
#Build the panel
xlimdat <- c(1, length(x1))
ylimdat <- c(-5, 5)
panel <- rp.control(title = "Eye-Ball-It", dat1=ylimdat, dat2=xlimdat, dat3=x1, dat4=x2, off=100, sca=1.0, size=c(300, 160))
rp.slider(panel, var=off, from=-500, to=500, action=lvm.draw, title="Offset", pos=c(5, 5, 290, 70), showvalue=TRUE)
rp.slider(panel, var=sca, from=0, to=2, action=lvm.draw, title="Scale", pos=c(5, 70, 290, 90), showvalue=TRUE)
It sounds like you want to minimize the function (Ax'+By) + (Az'+Bx) + (Ay'+Bz) for a pair of values: Namely, the time-offset: t0 and a time scale factor: tr. where Ax' = tr*(Ax + t0), etc..
I would look into SciPy's bivariate optimize functions. And I would use a mask or temporarily zero the data (both Ax' and By for example) over the "gaps" (assuming the gaps can be programmatically determined).
To make the process more efficient, start with a coarse sampling of A and B, but set the precision in fmin (or whatever optimizer you've selected) that is commensurate with your sampling. Then proceed with progressively finer-sampled windows of the full dataset until your windows are narrow and are not down-sampled.
Edit - matching axes
Regarding the issue of trying to identify which axis is co-linear with a given axis, and not knowing at thing about the characteristics of your data, i can point towards a similar question. Look into pHash or any of the other methods outlined in this post to help identify similar waveforms.

Categories

Resources