我正在尝试读取图像并将其转换为 python 中的双精度图像，我正在使用 cv2 库

Question

I tried doing the same thing in Matlab by reading the image using imread and then converting it to type double using im2double, how can I achieve the same thing in python or cv2 library?.

I tried using

img=imread('pathofmyimage')
double=np.asarray(img,dtype=np.float64)

but I am not sure of this method

Answer 1

You could just say

float_img = img.astype(np.float64)

But your way works too. The result from both methods will be equal.

To find out the type of an array, look at its .dtype . imread() returns np.uint8 by default, with values ranging 0 to 255, but it can return arrays of other types, if given the right flag. That is detailed in the docs.

After conversion with astype (or asarray ), the values will still be in the range of 0 to 255, just in a different data type. If you need them to be ranged 0.0 to 1.0, you'd need to scale the array (divide/multiply).