一、读入图片获得像素点的像素值、改变像素值、改变单个通道像素值、获得图像的行、列、图像数据类型、像素点、ROI区域.
import cv2
import numpy as np
#读取一个彩色图像
img = cv2.imread('C:/Users/NWPU/Desktop/1.jpg')
#cv2.imshow('image',img)
#获得某个像素点的像素值
px = img[200,200]
print(px)
#只获取绿色通道的像素值
green = img[200,200,1]
print(green)
#修改像素值
img[200,200] = [255,255,255]
print(img[200,200])
#使用Numpy数组的处理方法更好的获取像素点的值和编辑像素点的值
#获得指定像素点的红色通道的值
img_r = img.item(100,100,2)
print(img_r)
#修改指定像素点的红色通道的值
img.itemset((100,100,2),100)
img_rnew = img.item(100,100,2)
print(img_rnew)
#获取图片的信息:行数、列数、通道数、图像数据类型、像素数等
#获得图片的形状
print(img.shape) #(768,1024,3):768*1024大小的图像,彩色图像三通道
#查询像素总数
print(img.size)
#图片的数据类型,img.dtype在调试过程中很重要,因为很多opencv+python代码中的问题都是不合法的数据类型造成的
print(img.dtype) #实验图片为uint8数据类型
#图片的ROI:获得原始图片的一部分,将此部分复制到图片的另一个指定区域
img_ROI = img[280:340, 330:390]
img[273:333, 100:160] = img_ROI
cv2.rectangle(img,(280,330),(340,390),(255,255,255),1)
cv2.rectangle(img,(273,100),(333,160),(255,255,255),1)
cv2.imshow('image1',img)
#OpenCV存储彩色图片的格式是BGR模式,下面进行通道分离和合并
#使用split()函数进行通道分离,很耗时
b,g,r = cv2.split(img)
#使用merge()函数进行通道合并
img = cv2.merge((b,g,r))
#也可以直接操作Numpy数组来达到这一目的
b = np.zeros((img.shape[0],img.shape[1]),dtype = img.dtype)
g = np.zeros((img.shape[0],img.shape[1]),dtype = img.dtype)
r = np.zeros((img.shape[0],img.shape[1]),dtype = img.dtype)
b[:,:] = img[:,:,0]
g[:,:] = img[:,:,1]
r[:,:] = img[:,:,2]
运行结果:
二、图像相加
img1:
img2:
import cv2
import numpy as np
img1 = cv2.imread("C:/Users/NWPU/Desktop/1.jpg")
img2 = cv2.imread("C:/Users/NWPU/Desktop/2.jpg")
#图像相加:cv2.add()函数
rows, cols = img2.shape[:2] #获取img2的高度和宽度
img1_roi = img1[100:rows+100,100:cols+100]
img_plus = cv2.add(img1_roi,img2)
img1_copy_plus = img1.copy()
img1_copy_plus[100:rows+100,100:cols+100] = img_plus
cv2.imshow('img_plus',img1_copy_plus)
cv2.waitKey(0)
三、图像混合
import cv2
import numpy as np
img1 = cv2.imread("C:/Users/NWPU/Desktop/1.jpg")
img2 = cv2.imread("C:/Users/NWPU/Desktop/2.jpg")
#图像混合:cv2.addWeighted()函数
rows, cols = img2.shape[:2] #获取img2的高度和宽度
img1_roi = img1[100:rows+100, 100:cols+100]
img_mix = cv2.addWeighted(img1_roi, 0.3, img2, 0.7, 0)
img1_copy = img1.copy()
img1_copy[100:rows+100, 100:cols+100] = img_mix
cv2.imshow('img_mix',img1_copy)
cv2.waitKey(0)
结果:
四、图像的位运算
#图像的位操作有与、或、非、异或操作
'''
cv2.bitwise_and
cv2.bitwise_or
cv2.bitwsie_not
cv2.bitwise_xor
'''
import cv2
import numpy as np
img1 = cv2.imread("C:/Users/NWPU/Desktop/1.jpg") #768*1024
img2 = cv2.imread("C:/Users/NWPU/Desktop/2.jpg") #300*450
rows, cols = img2.shape[:2]
img1_roi = img1[100:rows+100, 100:cols+100]
img1_copy = img1.copy()
#与运算
img_add = cv2.bitwise_and(img1_roi,img2)
img1_copy[100:rows+100, 100:cols+100] = img_add
cv2.imshow('img_add', img1_copy)
#或运算
img_or = cv2.bitwise_or(img1_roi,img2)
img1_copy[100:rows+100, 100:cols+100] = img_or
cv2.imshow('img_or', img1_copy)
#非运算
img_not = cv2.bitwise_not(img1_roi,img2)
img1_copy[100:rows+100, 100:cols+100] = img_not
cv2.imshow('img_not', img1_copy)
#异或运算
img_xor = cv2.bitwise_xor(img1_roi,img2)
img1_copy[100:rows+100, 100:cols+100] = img_xor
cv2.imshow('img_xor', img1_copy)
cv2.waitKey(0)
结果:
与:
或:
非:
异或: