Scikit-learn: ValueError: unable to convert object to float64

Question

I had a very strange problem with scikit-image. Namely, I cannot run the "running example":

from skimage import data, io, filter

image = data.coins() # or any NumPy array!
edges = filter.sobel(image)
io.imshow(edges)

After that, Python 2.7 reports an error:

ValueError: can not convert object to float64.

In addition, I cannot open any "tif, png" images because Python reports a similar error. Interestingly, this problem arose suddenly, because at first everything worked correctly. Suppose this is due to the installation of the pymorph module, which probably affected numpy. However, do you have any ideas to solve this problem?