Scipy的几个简单图像操作

%matplotlib inline

Some simple usage of Scipy.ndimage

Bluring image

from PIL import Image
import numpy as np
from scipy.ndimage import filters
from matplotlib import pyplot as plt

blur gray image

im = np.array(Image.open(u'测试图片集/1.jpg').convert('L'))
im2 = filters.gaussian_filter(im,3)
im3 = filters.gaussian_filter(im,5)

plt.subplot(1,3,1)
plt.axis('off')
plt.imshow(im,cmap='gray')
plt.title('original')

plt.subplot(1,3,2)
plt.axis('off')
plt.imshow(im2,cmap='gray')
plt.title('gaussian(kernel 3)')

plt.subplot(1,3,3)
plt.axis('off')
plt.imshow(im3,cmap='gray')
plt.title('gaussian(kernel 5)')

blur color image

To blur color images, simply apply filter to each color channel

im = np.array(Image.open(u'测试图片集/1.jpg'))
im2 = np.zeros(im.shape)
im3 = np.zeros(im.shape)
for i in range(3):
    im2[:,:,i] = filters.gaussian_filter(im[:,:,i],3)
    im3[:,:,i] = filters.gaussian_filter(im[:,:,i],5)
im2 = np.uint8(im2)#color
im3 = np.uint8(im3)

plt.subplot(1,3,1)
plt.axis('off')
plt.imshow(im)
plt.title('original')

plt.subplot(1,3,2)
plt.axis('off')
plt.imshow(im2)
plt.title('gaussian(kernel 3)')

plt.subplot(1,3,3)
plt.axis('off')
plt.imshow(im3)
plt.title('gaussian(kernel 5)')

Image derivatives

Sobel filter

filters.sobel has 3 parament: input, direction, output

im = np.array(Image.open(u'测试图片集/1.jpg').convert('L'))

imx = np.zeros(im.shape)
filters.sobel(im,1,imx)

imy = np.zeros(im.shape)
filters.sobel(im,0,imy)

magnitude = np.sqrt(imx**2+imy**2)


plt.subplot(1,4,1)
plt.axis('off')
plt.imshow(im,cmap='gray')
plt.title('original')

plt.subplot(1,4,2)
plt.axis('off')
plt.imshow(imx,cmap='gray')
plt.title('x')

plt.subplot(1,4,3)
plt.axis('off')
plt.imshow(imy,cmap='gray')
plt.title('y')

plt.subplot(1,4,4)
plt.axis('off')
plt.imshow(magnitude,cmap='gray')
plt.title('x**2+y**2')

Due to background on book is white, I invert it.

plt.imshow(255.0-magnitude,cmap='gray')

Gaussian blur

for sigma in [2,5,10]: # standard deviation

    im = np.array(Image.open(u'测试图片集/1.jpg').convert('L'))

    imx = np.zeros(im.shape)
    filters.gaussian_filter(im,(sigma,sigma),(0,1),imx)

    imy = np.zeros(im.shape)

    filters.gaussian_filter(im,(sigma,sigma),(1,0),imy)
    magnitude = np.sqrt(imx**2+imy**2)

    plt.figure()
    plt.subplot(1,4,1)
    plt.axis('off')
    plt.imshow(im,cmap='gray')
    plt.title('original')

    plt.subplot(1,4,2)
    plt.axis('off')
    plt.imshow(imx,cmap='gray')
    plt.title('x (%d)'%sigma)

    plt.subplot(1,4,3)
    plt.axis('off')
    plt.imshow(imy,cmap='gray')
    plt.title('y (%d)'%sigma)

    plt.subplot(1,4,4)
    plt.axis('off')
    plt.imshow(magnitude,cmap='gray')
    plt.title('x**2+y**2 (%d)'%sigma)

    plt.show()

Morphology

from scipy.ndimage import morphology,measurements

im = np.array(Image.open(u'测试图片集/1.jpg').convert('L'))
im = 1*(im<128)

labels,nbr_object = measurements.label(im)
print 'Number of object:',nbr_object

im_open = morphology.binary_opening(im,np.ones((9,5)),iterations=2)#binary open???

labels_open,nbr_object_open = measurements.label(im_open)
print 'Number of object:',nbr_object_open

im_close = morphology.binary_closing(im,np.ones((9,5)),iterations=2)

labels_close,nbr_object_close = measurements.label(im_close)
print 'Number of object:',nbr_object_close

Number of object: 459
Number of object: 15
Number of object: 41

plt.subplot(1,3,1)
plt.axis('off')
plt.imshow(labels)
plt.title('raw binary')

plt.subplot(1,3,2)
plt.axis('off')
plt.imshow(labels_open)
plt.title('binary open')

plt.subplot(1,3,3)
plt.axis('off')
plt.imshow(labels_close)
plt.title('binary close')