3. 课后练习-图像处理

课后练习2

Required libs:Numpy PIL Scipy Matplotlib cv2 ## Q1. Write a python script to open the “lena.png” file using opencv. - Display the opened image in a new window named “Display Lena” - Save the image to a new file named “lena_resaved.png”

import cv2 as cv
img = cv.imread("lena.png") # cv2.imread('path')  read the img
cv.imshow("Display Lena",img) #cv2.imshow(windowname,path)
cv.waitKey(0) #to let the window display until clicking/pressing
cv.imwrite("lena_resaved.png",img) #cv2.imwrite(filename,path,params)

Q2. Use PIL and Matplotlib libraries for Q2.

Use “lena.png” to perform following operations and save the images:
- Crop a section from the image whose vertices are (100,100), (100,400), (400,100), (400,400).
(hint: convert the cv2 image into PIL Image)
- Rotate the cropped image by 45 degrees counter-clockwise. - Perform histogram equalization on lena.png. (hint: use ImageOps.equalize from PIL) - Use matplotlib to plot the histogram figure for both original image and processed image.
(hint: use histogram() function in PIL)
- Perform Max Filtering, Min Filtering, and Median Filter on lena.png. (hint: PIL.ImageFilter)
- Perform Gaussian Blur with sigma equal to 3 and 5. (hint: PIL.ImageFilter)

import cv2
from PIL import Image 
from PIL import ImageOps 
from PIL import ImageFilter as filter
from matplotlib import pyplot as plt  

pil_img=Image.open("lena.png") #open img in pil 
#(in cv2 lib, img is opened as array)
# load cv img: Image.fromarray()
pil_img.show() # show the img
img_crop = pil_img.crop((100,100,300,300)) #crop((start point,hight,width))
img_crop.show() #show the img

img_rota = img_crop.rotate(45) #rotate(degree)
img_rota.show()

img_eql=ImageOps.equalize(pil_img) 
#ImageOps.equalize(path) histogram equalize the imge

plt.plot(range(0,256),img_eql.histogram()) 
#pyplot(aix,img) plot someting   
#img.histogram()  return the histogram
plt.show()
plt.show(rang(0,256),pil_img.histogram())
plt.show()  

img_max = pil_img.filter(filter.maxfilter()) 
#filter.(Imagefilter.parm()) add filters
img_max.show()

img_min=pil_img.filter(filter.minfilter())
img_min.show()

img_mid=pil_img.filter(filter.medianfilter())
img_mid.show()


img_gus3=pil_img.filter(filter.gaussianblur(radius=3))
img_gus3.show()

img_gus10=pil_img.filter(filter.gaussianblur(radius=10))
img_gus10.show()

Q3. Colour space conversion. Use Python OpenCV functions to perform following operations on

“bee.png” and save the images at each step. - Read the image. - Convert the image to HSV(Hu Satuation Value:包含了三个通道：单色(H)，饱和度(S)，灰度(V)) color space. - Perform histogram equalization on V channel by cv2.equalizeHist(). - Convert the result image to BGR color space. - Show the image by cv2.imshow() and save the image.

import cv2
from PIL import ImageOps 
from PIL import ImageFilter as filter

bee_img = cv2.imread("bee.png")
bee_hsv = cv2.cvtColor(bee_img,cv2.COLOR_BGR2HSV)
bee_hsv.imshow()

bee_hsv[:,:,2]= cv2.equalizeHist(bee_hsv[:,:,2])
# 2 presents the channel 2: V
bee_rgb = cv2.cvtColor(bee_hsv,cv2.COLOR_HSV2BGR)
cv2.imshow("norm",bee_rgb)

bee_img[:,:,2]= cv2.equalizeHist(bee_rgb[:,:,2])
bee_img = cv2.cvtColor(bee_hsv,cv2.COLOR_HSV2BGR)
cv2.imshow("rgb",bee_img)