Python图像数据基本概念及操作

数据链接:https://pan.baidu.com/s/1cz8SihL2HYh_cFudc7y07Q 密码:tr35
1、图像数据基本操作

import cv2
import matplotlib.pyplot as plt
import numpy as np
from matplotlib import pyplot as plt
# 随机生成500x500的多维数组
random_image = np.random.random([500, 500])
# print(random_image)
plt.imshow(random_image, cmap='gray')
plt.colorbar()
# 加载灰度图像数据
img_gray_data = cv2.imread('./images/messi.jpg', cv2.IMREAD_GRAYSCALE)
print('数据类型:', type(img_gray_data))
print('数组类型:', img_gray_data.dtype)
print('数组形状：', img_gray_data.shape)
print('数组最大值：{}，最小值：{}'.format(img_gray_data.max(), img_gray_data.min()))
# 加载RGB图像数据
img_rgb_data = cv2.imread('./images/messi.jpg', cv2.IMREAD_COLOR)
print('数据类型:', type(img_rgb_data))
print('数组类型:', img_rgb_data.dtype)
print('数组形状：', img_rgb_data.shape)
print('数组最大值：{}，最小值：{}'.format(img_rgb_data.max(), img_rgb_data.min()))

2、显示图像

# 显示图像
cv2.imshow('gray', img_gray_data)
cv2.imshow('rgb', img_rgb_data)
cv2.waitKey(0)
cv2.destroyAllWindows()

3、像素值及访问操作

# 通过切片操作只显示某个通道上的图像数据
# OpenCV读取的图像数据通道排序为 BGR
cv2.imshow('B', img_rgb_data[:, :, 0])  # 颜色越接近蓝色，数值越大，亮度越亮
cv2.imshow('G', img_rgb_data[:, :, 1])
cv2.imshow('R', img_rgb_data[:, :, 2])
cv2.waitKey(0)
cv2.destroyAllWindows()
# 使用split进行通道分割
b_data, g_data, r_data = cv2.split(img_rgb_data)
cv2.imshow('B', b_data)
cv2.imshow('G', g_data)
cv2.imshow('R', r_data)
cv2.waitKey(0)
cv2.destroyAllWindows()
# 合并通道
bgr_data = cv2.merge((b_data, g_data, r_data))
cv2.imshow('color', bgr_data)
cv2.waitKey(0)
cv2.destroyAllWindows()
# 在图片上叠加一个蓝色方块
img_rgb_data[10:110, 10:110, :] = [255, 0, 0]
cv2.imshow('color with red box', img_rgb_data)
cv2.waitKey(0)
cv2.destroyAllWindows()
# ROI
ball = img_rgb_data[280:340, 330:390, :]
cv2.imshow('ball', ball)
cv2.waitKey(0)
cv2.destroyAllWindows()

4、图像融合

# NumPy加法和OpenCV中加法的区别
x = np.uint8([250])
y = np.uint8([10])
print(cv2.add(x, y))  # 250 + 10 = 260 => 255
print(x + y)          # 250 + 10 = 260 % 256 = 4
img1 = cv2.imread('./images/python_ai.png')
img2 = cv2.imread('./images/python_logo.png')
img2 = cv2.resize(img2, (img1.shape[1], img1.shape[0]))
dst = cv2.addWeighted(img1, 0.7, img2, 0.3, 0)
cv2.imshow('dst', dst)
cv2.waitKey(0)
cv2.destroyAllWindows()
cv2.imwrite('./output/python_ai_logo.png', dst)

5、色彩空间

img_gray_data = cv2.imread('./images/messi.jpg', cv2.IMREAD_GRAYSCALE)
img_bgr_data = cv2.cvtColor(img_gray_data, cv2.COLOR_GRAY2BGR)
cv2.imshow('gray', img_gray_data)
cv2.imshow('bgr', img_bgr_data)
cv2.waitKey(0)
cv2.destroyAllWindows()
img_bgr_data = cv2.imread('./images/messi.jpg')
img_hsv_data = cv2.cvtColor(img_bgr_data, cv2.COLOR_BGR2HSV)
cv2.imshow('bgr', img_bgr_data)
cv2.imshow('hsv', img_hsv_data)
cv2.waitKey(0)
cv2.destroyAllWindows()

6、颜色直方图

img_gray_data = cv2.imread('./images/messi.jpg', cv2.IMREAD_GRAYSCALE)
hist, bins = np.histogram(img_gray_data.ravel(), bins=10)
print(hist)
print(bins)
plt.hist(img_gray_data.ravel(), bins=50);
# 彩色图像直方图
img_bgr_data = cv2.imread('./images/messi.jpg')
plt.figure(figsize=(15, 5))
# B通道 直方图
ax1 = plt.subplot(131)
ax1.hist(img_bgr_data[:, :, 0].ravel(), bins=50, color='b')
# G通道 直方图
ax2 = plt.subplot(132)
ax2.hist(img_bgr_data[:, :, 1].ravel(), bins=50, color='g')
# # B通道 直方图
ax3 = plt.subplot(133)
ax3.hist(img_bgr_data[:, :, 2].ravel(), bins=50, color='r')
# 直方图均衡化
wiki_img = cv2.imread('./images/wiki.jpg', cv2.IMREAD_GRAYSCALE)
cv2.imshow('wiki_img', wiki_img)
cv2.waitKey(0)
cv2.destroyAllWindows()
plt.hist(wiki_img.ravel(), bins=256, range=[0, 256])
plt.show()

equ_wiki_img = cv2.equalizeHist(wiki_img)
cv2.imshow('equ_wiki_img', equ_wiki_img)
cv2.waitKey(0)
cv2.destroyAllWindows()

plt.hist(equ_wiki_img.ravel(), bins=256, range=[0, 256])
plt.show()

7、图像滤波

# 7.1 中值滤波
img = cv2.imread('./images/nosiy_image.jpg', cv2.IMREAD_GRAYSCALE)
med1 = cv2.medianBlur(img, 3) # 3x3中值滤波
med2 =cv2.medianBlur(img, 5) # 5x5中值滤波

# 图像对比 
cv2.imshow('img', img)
cv2.imshow('med1', med1)
cv2.imshow('med2', med2)
cv2.waitKey(0)
cv2.destroyAllWindows()
print(img.dtype)
# 7.2 高斯滤波
img = cv2.imread('./images/camera.jpg', cv2.IMREAD_GRAYSCALE)
gas1 = cv2.GaussianBlur(img, (3, 3), 0) # 3x3
gas2 = cv2.GaussianBlur(img, (5, 5), 0) # 5x5
# 图像对比 
cv2.imshow('img', img)
cv2.imshow('gas1', gas1)
cv2.imshow('gas2', gas2)
cv2.waitKey(0)
cv2.destroyAllWindows()
# 7.3 均值滤波
img = cv2.imread('./images/camera.jpg', cv2.IMREAD_GRAYSCALE)
mean1 = cv2.blur(img, (3, 3), 0) # 3x3
mean2 = cv2.blur(img, (5, 5), 0) # 5x5
# 图像对比 
cv2.imshow('img', img)
cv2.imshow('mean1', mean1)
cv2.imshow('mean2', mean2)
cv2.waitKey(0)
cv2.destroyAllWindows()

8、边缘检测

img = cv2.imread('./images/messi.jpg', cv2.IMREAD_GRAYSCALE)
laplacian = cv2.Laplacian(img, cv2.CV_64F)
sobelx = cv2.Sobel(img, cv2.CV_64F, 1, 0, ksize=5)
sobely = cv2.Sobel(img, cv2.CV_64F, 0, 1, ksize=5)
# 图像对比 
plt.figure(figsize=(10, 8))
plt.subplot(2,2,1),plt.imshow(img,cmap = 'gray')
plt.title('Original'), plt.xticks([]), plt.yticks([])
plt.subplot(2,2,2),plt.imshow(laplacian,cmap = 'gray')
plt.title('Laplacian'), plt.xticks([]), plt.yticks([])
plt.subplot(2,2,3),plt.imshow(sobelx,cmap = 'gray')
plt.title('Sobel X'), plt.xticks([]), plt.yticks([])
plt.subplot(2,2,4),plt.imshow(sobely,cmap = 'gray')
plt.title('Sobel Y'), plt.xticks([]), plt.yticks([])
plt.show()

Python图像数据基本概念及操作

猜你喜欢