Day 2 - Probability, OpenCV_matplotlib

Probability

sample(오메가) : 실험에서 나올 수 있는 모든 결과

random variable : 어떤 사건에 대해 나올 수 있는 값을 결정하는 함수(function)

Event Space: 샘플 셋이 나올 수 있는 power set

Bayes' Theorem

posterior ∝ linklihood × prior

Gaussian Distribution (정규 분포)

평균과 분산을 가지고 estimate을 할 수 있다.

OpenCV_matplotlib

OpenCV

컴퓨터 비전을 목적으로 하는 오픈 소스 라이브러리

pip install opencv-python

https://docs.opencv.org/4.x/index.html

[OpenCV: OpenCV modules

OpenCV 4.10.0-dev Open Source Computer Vision

docs.opencv.org](https://docs.opencv.org/4.x/index.html)

Matpoltlib

python과 numpy array를 기반으로 plotting과 시각화를 목적으로 하는 라이브러리

pip install matplotlib

https://matplotlib.org/stable/index.html

[Matplotlib documentation — Matplotlib 3.9.2 documentation

Matplotlib 3.9.2 documentation Matplotlib is a comprehensive library for creating static, animated, and interactive visualizations. Install For more detailed instructions, see the installation guide. Learn What's new Learn about new features and API change

matplotlib.org](https://matplotlib.org/stable/index.html)

Image Read and Visualize

import cv2
import matplotlib.pyplot as plt

im_name='img1.jpg'
image=cv2.imread(im_name)
image

// 이미지 나타내기
plt.imshow(image)

Multiple Image Visualize, figure/axs in matplotlib

// 이미지  rgb로 변환
image_rgb=cv2.cvtColor(image,cv2.COLOR_BGR2RGB)
plt.imshow(image_rgb)

// Grayscale로 이미지 나타내기
image_gray=cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
plt.imshow(image_gray,cmap='gray')

// 세 장의 이미지 병렬로 표현하기
im_name='img1.jpg'
image=cv2.imread(im_name)
image=cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
image_rgb=cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
image_gray=cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)

plt.figure(figsize(15,8))

plt.subplot(1,3,1)
plt.imshow(image)
plt.axis('off')
plt.title('Image 1')

plt.subplot(1,3,2)
plt.imshow(image_rgb)
plt.axis('off')
plt.title('Image 2')

plt.subplot(1,3,3)
plt.imshow(image_gray)
plt.axis('off')

plt.show()

라인 플롯

import matplotlib.pyplot as plt
import numpy as np
a=np.linespace(0,10,100)
b=np.exp(-a)
plt.plot(a,b)
plt.show()

히스토그램

import matplotlib.pyploy as plt
from numpy.random import normal,rand
x=normal(size=200)
plt.hist(x,bins=30)
plt.show()

산점도

import matplotlib.pyplot as plt
from numpy.random import rand
a=rand(100)
b=rand(100)
plt.scatter(a,b)
plt.show()

3D 플롯

from matplotlib import cm
from mpl_toolkits.mplot3d import Axes3D
import matplotlib.pyplot as plt
import numpy as np
fig=plt.figure()
ax=fig.gca(projection='3d')
X=np.arange(-5,5,0.25)
Y=np.arange(-5,5,0.25)
X,Y=np.meshgrid(X,Y)
R=np.sqrt(X**2+Y**2)
Z=np.sin(R)
surf=ax.plot_surface(X,Y,Z,rstride=1,cstride=1,cmap=cm.coolwarm)
plt.show()

Image Cropping, masking and save

import cv2
import matplotlib.pyplot as plt
def read_image(im_name):
	image=cv2.imread(im_name)
    image_rgb=cv2.cvtColor(image,cv2.COLOR_BGR2RGB)
    plt.imshow(image_rgb)
    plt.show()
    return image_rgb
    
 ## To make function and read image2
 im1=read_image('image.jpeg')
 im2=read_image('image2.jpeg')
 
 print(im1.shape,im2.shape)

image cropping

h1,w1=im1.shape[:2]
h2,w2=im2.shape[:2]
size=720

def center_crop(im,h,w,size=720):
	im=im[int(h/2-size/2):int(h/2+size/2), int(w/2-size/2):int(w/2+size/2)]
    return im
    
im1=center_crop(im1,h1,w1,size);
im2=center_crop(im2,h2,w2,size)

def two_image_show(im1,im2):
	plt.subplot(1,2,1)
    plt.imshow(im1)
    plt.subplot(1,2,2)
    plt.imshow(im2)
    plt.show()
    
two_image_show(im1,im2)

image masking

import numpy as np

im1_mask=np.zeros_like(im1)
im1_mask[int(size/2:,]=im1[int(size/2):,]
im2_mask=np.zeros_like(im2)
im2_mask[:int(size/2),]=im2[:int(size/2),]

two_image_show(im1_mask,im2_mask)

merge=cv1.bitwise_or(im1_mask,im2_mask)
plt.imshow(merge)
plt.show()

circle mask

circle_mask=np.zeros_like(im2)
# cv2.circle(대상이미지,(원점x,원점y),반지름,(색상),채우기)
cv2.circle(circle_mask,(size//2,size//2),size//3,(255,255,255),-1)

masked)im2=cv2.bitwise_and(im2,circle_mask)

plt.imshow(masked_im2)

저장하기

plt.imsave('save_img.plg',masked_im2)

Image matching and visualize

import numpy as np
import cv2 as cv
import matplotlib.pyplot as plt

queryImage='cv_cover.jpg'
trainImage='cv_desk.png'

plt.figure(figsize=(15,8))
plt.subplot(1,2,1)
plt.imshow(cv.cvtColor(cv.imread(queryImage),cv.COLOR_BGR2RGB))
plt.title('QueryImage')
plt.subplot(1,2,2)
plt.imshow(cv.cvtColor(cv.imread(trainImage),cv.COLOR_BGR2RGB))
plt.title('Dest image')
plt.show()

img1=cv.imread(queryImage,cv.IMREAD_GRAYSCALE)
img2=cv.imread(trainImage,cv.IMREAD_GRAYSCALE)

# Initiate IRB detector
orb=cv.ORB_create()

# find the keypoints and descriptiors with ORB
kp1,des1=orb.detoectAndConpute(img1,None)
kp2,des2=orb.detectAndCompute(img2,None)

# create BFMatcher object
bf=cv.BFMatcher(cv.NORM_HAMMING,crossCheck=True)

#matches=bf.match(des1,des2)

# Sort them in the order of their distance
matches=sorted(matches,key=lanbda x:x.distance)

# Draw first 10 matches
img3=cv.drawMatches(img1,kp1,img2,kp2,matches[:10],None,flags=cv.DrawMatchesFlags_NOT_DRAW_SINGLE_POINT

plt.figure(figsize=(15,8))
plt.imshow(img3)
plt.show()

Find homography and warping

# Extract location of good matches
points1=np.zeros((len(matches),2),dtype=np.float32)
points2=np.zeros((len(matches),2),dtype=np.float32)

for i,match in enumerate(matches):
	points1[i,:]=kp1[match.queryIdx].pt
	points2[i,:]=kp2[match.trainIdx].pt
    
h,mask=cv.findHomography(points1,points2,cv.RANSAC)

height,width=img2.shape

im1_warped=cv.warpPerspective(img1,h,(width,height)


plt.figure(figsize=(15,8))
plt.subplot(1,3,1)
plt.imshow(img1,cmap='gray')
plt.subplot(1,3,2)
plt.imshow(img2,cmap='gray')
plt.subplot(1,3,3)
plt.imshow(img1_warped,cmap='gray')
plt.show()

Brute-Force matching with SIFT Descriptors and Ratio Test

# Initiate SIFT detector
sift=cv.SIFT_create()

# find the keypoints and descriptos with SIFT
kp1,des1=sift.detectAndCompute(img1,None)
kp2,des2=sift.detectAndCompute(img2,None)

# BFMatcher with default params
bf=cv.BFMatcher()
matches=bf.knnMatch(des1,des2,k=2)

# Apply ratio test
good=[]
for m,n in matches:
	if m.distance < 0.75*n.distance:
    	good.append([m])
        
# cv.drawMatchesKnn expects list of lists as matches
img3=cv.drawMatchesKnn(img1,kp1,img2,kp2,good,None,flags=cv.DrawMatchesFlags_NOT_DRAW_SINGLE_POINT

plt.figure(figsize=(15,8))
plt.imshow(img3)
plt.show()