[英]How to transform a FFT (Fast Fourier Transform) into a Polar Transformation with Python?
I was able to create a FFT Transformation from my image but I don't know how to continue... 我可以根据自己的图像创建FFT转换,但不知道如何继续...
I am using this to solve my problem: Align text for OCR 我正在使用它来解决我的问题: 对齐OCR的文本
This code was all that worked for me until now: 到目前为止,这段代码对我一直有效:
import cv2
import numpy as np
from matplotlib import pyplot as plt
%matplotlib inline
img = cv2.imread(r'test.jpg', cv2.IMREAD_GRAYSCALE)
f = np.fft.fft2(img)
fshift = np.fft.fftshift(f)
magnitude_spectrum = 20 * np.log(np.abs(fshift))
plt.subplot(121), plt.imshow(img, cmap='gray')
plt.title('Input Image'), plt.xticks([]), plt.yticks([])
plt.subplot(122), plt.imshow(magnitude_spectrum, cmap='gray')
plt.title('Magnitude Spectrum'), plt.xticks([]), plt.yticks([])
plt.show()
I need the mean value generated from a Polar Transformation, but I don't know how to transform a FFT to a Polar Transformation in Python . 我需要从Polar转换生成的平均值,但是我不知道如何在Python中将FFT转换为Polar转换 。
This is roughly solution to you problem; 这大致解决了您的问题; It was tested on one sample image, and the result looks credible. 在一个样本图像上进行了测试,结果看起来是可信的。
# your code goes here...
def transform_data(m):
dpix, dpiy = m.shape
x_c, y_c = np.unravel_index(np.argmax(m), m.shape)
angles = np.linspace(0, np.pi*2, min(dpix, dpiy))
mrc = min(abs(x_c - dpix), abs(y_c - dpiy), x_c, y_c)
radiuses = np.linspace(0, mrc, max(dpix, dpiy))
A, R = np.meshgrid(angles, radiuses)
X = R * np.cos(A)
Y = R * np.sin(A)
return A, R, m[X.astype(int) + mrc - 1, Y.astype(int) + mrc - 1]
angles, radiuses, m = transform_data(magnitude_spectrum)
plt.contourf(angles, radiuses, m)
Finally, we can get the angle we want to turn the original image: 最后,我们可以得到想要旋转原始图像的角度:
sample_angles = np.linspace(0, 2 * np.pi, len(c.sum(axis=0))) / np.pi*180
turn_angle_in_degrees = 90 - sample_angles[np.argmax(c.sum(axis=0))]
For my sample image I got: 对于我的示例图像,我得到了:
turn_angle_in_degrees = 3.2015810276679844 degrees.
Also, we can plot projected spectrum magnitude: 此外,我们可以绘制预计的频谱幅度:
plt.plot(sample_angles, c.sum(axis=0))
Hope that helps... 希望有帮助...
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.