dtype not recognized/unsupported

[nightrook779]

I am following the example in tutorial.
I am getting a dtype not recognized/unsupported exception at

bkg_image = bkg.back()

when I execute

bkg = sep.Background(data)
bkg_image = bkg.back()

What am I doing wrong?

[kbarbary]

What is the dtype of data? You can see this by doing data.dtype.

[nightrook779]

dtype of data:
<class 'numpy.ndarray'>

dtype of bkg:
<class 'sep.Background'>

[kbarbary]

You should see something like dtype('float64'). I'm not sure how you could get that result for data.dtype. Is data a numpy array? Maybe you could post a minimal example.

[nightrook779]

Oh. I see what happened.
I am using openCV to import and manipulate the image, since I can't use fitsio. My bad.
Any alternatives/thoughts on how to do this?

[kbarbary]

Regardless of the method you use to read the data, you need to get it in the form of a 2-d numpy array of floats or ints. This should be possible with opencv, but I'm not that familiar with the library.

[nightrook779]

Alright. I was able to get the code working. Whats happening is that since I am reading jpeg files, numpy array is uint8. Converting to float64 fixed the issue. Code I used:

data = cv2.imread('pinwheel.jpg')
data = cv2.cvtColor(data,cv2.COLOR_BGR2GRAY)
data = data.astype(np.float64)