使用Python OpenCV将图像捕获为数组

Question

I've seen from the sample how to display an image from the webcam, but how do I get the image captured as an array?

import cv

capture = cv.CaptureFromCAM(0)
img = cv.QueryFrame(capture)

img.tostring() gives me weird caracters. Thanks in adv.

Answer 1

I think this is what you're looking for:

img=cv.LoadImage("asd.png")
mat=cv.GetMat(img)
mat[3,1]
(83.0, 88.0, 89.0)

anyway, you should check opencv python cookbook for use with PIL and NUMPY packages.

Answer 2

Be sure to get an opencv with numpy support included. The sequence that works for me is capture an IPL, convert to cvMat, convert to numpy:

import cv, numpy, pylab
capture = cv.CaptureFromCAM(0)
img = cv.QueryFrame(capture)
mat=cv.GetMat(img)
a = numpy.asarray(mat)
pylab.imshow(a)

Note that the representation of color is different from what pylab assumes. But there comes opencv documentation to your help!

Answer 3

I believe your problem might be the way you are printing/displaying the contents of the array.

Nevertheless, this blog post shows how to capture frames from a webcam on linux using python.

Answer 4

According to the docs, QueryFrame returns an IplImage . IplImage can be treated as an arbitrary array . It looks like you'll want to use the "Get" set of functions to access array elements.