基于python的OpenCV图像处理学习笔记
打开.ipynb文件
- 打开anaconda prompt
- 输入jupyter notebook
- 找到文件打开即可
1. 图像简介与图像的基本操作
- 图像由很多的小方格组成（像素）
- 彩色图像每个像素有三个通道RGB,每个通道对应一个值（0-255）
import cv2 #opencv读取的格式是BGR
import matplotlib.pyplot as plt
import numpy as np
# 读取数据
img = cv2.imread('./image/cat.jpg')
# print(img)
# 图像的显示,也可以创建多个窗口
# cv2.imshow('image', img)
# # 等待时间，毫秒级，0表示任意键终止
# cv2.waitKey(0)
# cv2.destroyAllWindows()
# 图形的大小（像素）和通道
# print(img.shape)
# 灰度处理
# img=cv2.imread('./image/cat.jpg', cv2.IMREAD_GRAYSCALE)
# cv2.imshow('cat',img)
# cv2.waitKey(0)
# cv2.destroyAllWindows()
# print(img.shape)
#
# # 图像保存
# cv2.imwrite('./image/gray_cat.jpg', img)
#
# # 图像信息
# print(type(img)) # <class 'numpy.ndarray'>
# print(img.size) # (398, 600) = 238800
# print(img.dtype) # uint8
# 读取视频
# vc = cv2.VideoCapture('./image/test.mp4')
# # 检查是否打开正确
# if vc.isOpened():
# open, frame = vc.read()
# else:
# open = False
#
# # 能正常打开
# while open:
# ret, frame = vc.read()
# if frame is None:
# break
# if ret == True:
# gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
# cv2.imshow('result', gray)
# if cv2.waitKey(100) & 0xFF == 27: # 每隔0.1s刷新一次(越小播放越快) 退出键Esc退出
# break
# # 释放资源
# vc.release()
# cv2.destroyAllWindows()
# 截取部分图像数据
# 声明图形显示方法
def cv_show(name,img):
cv2.imshow(name,img)
cv2.waitKey(0)
cv2.destroyAllWindows()
# 只取到了50*200的图像部分
# cat=img[0:50,0:200]
# cv_show('cat',cat)
# 颜色通道提取
# b,g,r=cv2.split(img)
# print(b)
# print(b.shape)
# 合并图像
# img=cv2.merge((b,g,r))
# print(img.shape)
# 只保留R
# cur_img = img.copy()
# print(img)
#
# # bgr分别对应0,1,2 cur_img[:,:,0]的意思为任意行，任意行内的任意像素点（列），的第一个值（B）为0
# cur_img[:,:,0] = 0 #B
# cur_img[:,:,1] = 0 #G
# print(cur_img)
# cv_show('R',cur_img)
#
# # 只保留G
# cur_img = img.copy()
# cur_img[:,:,0] = 0
# cur_img[:,:,2] = 0
# cv_show('G',cur_img)
#
# # 只保留B
# cur_img = img.copy()
# cur_img[:,:,1] = 0
# cur_img[:,:,2] = 0
# cv_show('B',cur_img)
# 边界填充
# BORDER_REPLICATE：复制法，也就是复制最边缘像素。
# BORDER_REFLECT：反射法，对感兴趣的图像中的像素在两边进行复制例如：fedcba|abcdefgh|hgfedcb
# BORDER_REFLECT_101：反射法，也就是以最边缘像素为轴，对称，gfedcb|abcdefgh|gfedcba
# BORDER_WRAP：外包装法cdefgh|abcdefgh|abcdefg
# BORDER_CONSTANT：常量法，常数值填充。
# 设置填充的值
# top_size,bottom_size,left_size,right_size = (50,50,50,50)
# # 填充的边界模糊的
# replicate = cv2.copyMakeBorder(img, top_size, bottom_size, left_size, right_size, borderType=cv2.BORDER_REPLICATE)
# # 镜像填充
# reflect = cv2.copyMakeBorder(img, top_size, bottom_size, left_size, right_size,cv2.BORDER_REFLECT)
# # 稍加模糊的镜像填充
# reflect101 = cv2.copyMakeBorder(img, top_size, bottom_size, left_size, right_size, cv2.BORDER_REFLECT_101)
# wrap = cv2.copyMakeBorder(img, top_size, bottom_size, left_size, right_size, cv2.BORDER_WRAP)
# constant = cv2.copyMakeBorder(img, top_size, bottom_size, left_size, right_size,cv2.BORDER_CONSTANT, value=0)
# 数值计算
# img_cat=cv2.imread('./image/cat.jpg')
# img_dog=cv2.imread('./image/dog.jpg')
#
# img_cat2= img_cat +10
# print(img_cat[:5,:,0]) # 只显示5行，和B的值
# print(img_cat2[:5,:,0])
#
# # 0-255越界，相当于% 256
# print((img_cat + img_cat2)[:5,:,0] )
#
# # add则不会越界，最大就为255
# print(cv2.add(img_cat,img_cat2)[:5,:,0])
# 图像融合
# 融合的前提 两张图像的shape要一致
# img_cat=cv2.imread('./image/cat.jpg')
# img_dog=cv2.imread('./image/dog.jpg')
# print(img_cat.shape)
# print(img_dog.shape)
#
# # 修改shape
# img_dog = cv2.resize(img_dog, (600, 398))
# print(img_dog.shape)
#
# # 按照权重融合(权重越大越明显)
# res = cv2.addWeighted(img_cat, 0.2, img_dog, 0.8, 0)
# cv_show('res',res)
# 设置图像宽高比例
# 正方形(默认总高或宽为2000)
res = cv2.resize(img, (0, 0), fx=4, fy=4)
cv_show('res',res)# 2000*2000
res = cv2.resize(img, (0, 0), fx=2, fy=1)
cv_show('res',res) # 1000*500
res = cv2.resize(img, (500, 250), fx=2, fy=1)
cv_show('res',res) # 500*250
2.图像处理
# 图像处理
import cv2
import numpy as np
def cv_show(name,img):
cv2.imshow(name,img)
cv2.waitKey(0)
cv2.destroyAllWindows()
# 灰度图
img= cv2.imread('./image/cat.jpg')
img_gray = cv2.cvtColor(img,cv2.COLOR_RGB2GRAY)
# print(img_gray.shape)
# cv_show('img_gray',img_gray)
#
# #HSV
# # H - 色调（主波长）
# # S - 饱和度（纯度/颜色的阴影）。
# # V值（强度）
#
# hsv = cv2.cvtColor(img,cv2.COLOR_RGB2HSV)
# cv_show('hsv',hsv)
# 图像阈值
# ret, dst = cv2.threshold(src, thresh, maxval, type)
# src： 输入图，只能输入单通道图像，通常来说为灰度图
# dst： 输出图
# thresh： 阈值
# maxval： 当像素值超过了阈值（或者小于阈值，根据type来决定），所赋予的值
# type：二值化操作的类型，包含以下5种类型： cv2.THRESH_BINARY； cv2.THRESH_BINARY_INV； cv2.THRESH_TRUNC； cv2.THRESH_TOZERO；cv2.THRESH_TOZERO_INV
# cv2.THRESH_BINARY 超过阈值部分取maxval（最大值），否则取0
# cv2.THRESH_BINARY_INV THRESH_BINARY的反转
# cv2.THRESH_TRUNC 大于阈值部分设为阈值，否则不变
# cv2.THRESH_TOZERO 大于阈值部分不改变，否则设为0
# cv2.THRESH_TOZERO_INV THRESH_TOZERO的反转
# 0为黑,255为白
# ret, thresh1 = cv2.threshold(img_gray, 127, 255, cv2.THRESH_BINARY)
# cv_show('thresh1',thresh1)
# ret, thresh2 = cv2.threshold(img_gray, 127, 255, cv2.THRESH_BINARY_INV)
# cv_show('thresh2',thresh2)
# ret, thresh3 = cv2.threshold(img_gray, 127, 255, cv2.THRESH_TRUNC)
# cv_show('thresh3',thresh3)
# ret, thresh4 = cv2.threshold(img_gray, 127, 255, cv2.THRESH_TOZERO)
# cv_show('thresh4',thresh4)
# ret, thresh5 = cv2.threshold(img_gray, 127, 255, cv2.THRESH_TOZERO_INV)
# cv_show('thresh5',thresh5)
# 图像滤波(平滑处理)
# 去除噪声
img = cv2.imread('./image/lenaNoise.png')
cv_show('noise',img)
# img = cv2.cvtColor(img,cv2.COLOR_RGB2GRAY)
#
# # 均值滤波
# # 矩阵全部相加除以个数,求平均
# print(img)
blur = cv2.blur(img,(3,3))
# cv_show('blur',blur)
# print('===============')
# print(blur)
# 方框滤波
# 基本和均值一样，可以选择归一化, -1 得到的结果和原始通道数表示一致,normalize归一化(求平均)
# box = cv2.boxFilter(img,-1,(3,3), normalize=True)
# cv_show('box',box)
#
#
# # 不求平均,全部加起来,越界值如果大于255,就设置为255
# box = cv2.boxFilter(img,-1,(3,3), normalize=False)
# cv_show('box',box)
# 高斯滤波
# 高斯模糊的卷积核里的数值是满足高斯分布，相当于更重视中间的(更重视相邻的(上下左右)),乘以权重
aussian = cv2.GaussianBlur(img, (3, 3), 1)
# cv_show('aussian',aussian)
# 中值滤波
# 相当于用中值代替
# 排序后取中间值,作为结果 5表示5*5
# 去噪声效果最好
median = cv2.medianBlur(img, 5) # 中值滤波
# cv_show('median', median)
# 展示所有的
# np.hstack横向显示,np.stack纵向显示
res = np.hstack((blur,aussian,median))
#print (res)
cv_show('median vs average', res)
3. 形态学基本操作
import cv2
import numpy as np
def cv_show(name,img):
cv2.imshow(name,img)
cv2.waitKey(0)
cv2.destroyAllWindows()
# 形态学-腐蚀操作
# 腐蚀就是将白的边缘变黑一点,kernel数组越大,腐蚀面积越大
img = cv2.imread('./image/dige.png')
# cv_show('dige',img)
#
# kernel = np.ones((3,3), np.uint8) # ones返回一个全1的n维数组,这里返回3*3,np.uint8为数据类型
# erosion = cv2.erode(img,kernel,iterations = 1) # iterations = 1表示迭代一次
# cv_show('erosion',erosion)
#
#
pie = cv2.imread('./image/pie.png')
# cv_show('pie',pie)
# # 验证迭代次数,迭代次数越多,白色的圆圈越小,并且不在是园形了
# kernel = np.ones((30,30),np.uint8)
# erosion_1 = cv2.erode(pie,kernel,iterations = 1)
# erosion_2 = cv2.erode(pie,kernel,iterations = 2)
# erosion_3 = cv2.erode(pie,kernel,iterations = 3)
# res = np.hstack((erosion_1,erosion_2,erosion_3))
# cv_show('res',res)
# 形态学-膨胀操作
# kernel = np.ones((3,3), np.uint8)
# dige_erosion = cv2.erode(img,kernel,iterations = 1)
# cv_show('dige_erosion',dige_erosion)
# # 膨胀则为将白色边缘扩大
# kernel = np.ones((3,3),np.uint8)
# dige_dilate = cv2.dilate(dige_erosion,kernel,iterations = 1)
# cv_show('dilate', dige_dilate)
#
# kernel = np.ones((30,30),np.uint8)
# dilate_1 = cv2.dilate(pie,kernel,iterations = 1)
# dilate_2 = cv2.dilate(pie,kernel,iterations = 2)
# dilate_3 = cv2.dilate(pie,kernel,iterations = 3)
# res = np.hstack((dilate_1,dilate_2,dilate_3))
# cv_show('res', res)
# 开运算
# 先进行腐蚀,再进行膨胀
# kernel = np.ones((3,3),np.uint8)
# opening = cv2.morphologyEx(img, cv2.MORPH_OPEN, kernel)
# cv_show('opening', opening)
#
# # 闭运算
# # 先进行膨胀,再进行腐蚀
# kernel = np.ones((3,3),np.uint8)
# closing = cv2.morphologyEx(img, cv2.MORPH_CLOSE, kernel)
# cv_show('closing', closing)
# 梯度运算
# 梯度 = 膨胀 - 腐蚀
# kernel = np.ones((7,7),np.uint8)
# dilate = cv2.dilate(pie,kernel,iterations = 5)
# erosion = cv2.erode(pie,kernel,iterations = 5)
# res = np.hstack((dilate,erosion))
# cv_show('res', res)
# # 剩下膨胀出来的部分
# gradient = cv2.morphologyEx(pie, cv2.MORPH_GRADIENT, kernel)
# cv_show('gradient', gradient)
# 礼帽和黑帽
# 礼帽 = 顶帽 = 原始输入-开运算结果
kernel = np.ones((7,7),np.uint8)
tophat = cv2.morphologyEx(img, cv2.MORPH_TOPHAT, kernel)
cv_show('tophat', tophat)
# 黑帽 = 闭运算-原始输入
blackhat = cv2.morphologyEx(img, cv2.MORPH_BLACKHAT, kernel)
cv_show('blackhat', blackhat)
4. 梯度处理
import cv2
import numpy as np
def cv_show(name, img):
cv2.imshow(name, img)
cv2.waitKey(0)
cv2.destroyAllWindows()
# 图像梯度-Sobel算子
# dst = cv2.Sobel(src, ddepth, dx, dy, ksize)
# ddepth:图像的深度（输入输出一样）
# dx和dy分别表示水平和竖直方向 dx :右边-左边，中间为0 dy:上边-下边，中间0
# ksize是Sobel算子的大小
# img = cv2.imread('./image/pie.png')
# cv_show('pie',img)
#
# sobelx = cv2.Sobel(img,cv2.CV_64F,1,0, ksize=3)
# sobelx = cv2.convertScaleAbs(sobelx)
# cv_show('sobelx',sobelx)
#
# sobely = cv2.Sobel(img,cv2.CV_64F,0,1,ksize=3)
# cv_show('sobely',sobely)
#
# sobelxy = cv2.addWeighted(sobelx,0.5,sobely,0.5,0)
# cv_show('sobelxy',sobelxy)
# 分开计算合并
# img = cv2.imread('./image/lena.jpg',cv2.IMREAD_GRAYSCALE)
# sobelx = cv2.Sobel(img,cv2.CV_64F,1,0,ksize=3)
# # 绝对值
# sobelx = cv2.convertScaleAbs(sobelx)
# sobely = cv2.Sobel(img,cv2.CV_64F,0,1,ksize=3)
# sobely = cv2.convertScaleAbs(sobely)
# sobelxy = cv2.addWeighted(sobelx,0.5,sobely,0.5,0)
# cv_show('sobelxy',sobelxy)
#
# # 直接计算效果更差
# sobelxy = cv2.Sobel(img,cv2.CV_64F,1,1,ksize=3)
# sobelxy = cv2.convertScaleAbs(sobelxy)
# cv_show('sobelxy',sobelxy)
# scharr算子
# 核变大了，结果更敏感一些
# laplacian算子
#0 1 0
#1 -4 1
#0 1 0
# 中间点与中间的的四周
#不同算子的差异
img = cv2.imread('./image/lena.jpg',cv2.IMREAD_GRAYSCALE)
# 边界
sobelx = cv2.Sobel(img,cv2.CV_64F,1,0,ksize=3)
sobely = cv2.Sobel(img,cv2.CV_64F,0,1,ksize=3)
sobelx = cv2.convertScaleAbs(sobelx)
sobely = cv2.convertScaleAbs(sobely)
sobelxy = cv2.addWeighted(sobelx,0.5,sobely,0.5,0)
# 更加敏感，更细致
scharrx = cv2.Scharr(img,cv2.CV_64F,1,0)
scharry = cv2.Scharr(img,cv2.CV_64F,0,1)
scharrx = cv2.convertScaleAbs(scharrx)
scharry = cv2.convertScaleAbs(scharry)
scharrxy = cv2.addWeighted(scharrx,0.5,scharry,0.5,0)
# 效果较差
laplacian = cv2.Laplacian(img,cv2.CV_64F)
laplacian = cv2.convertScaleAbs(laplacian)
res = np.hstack((sobelxy,scharrxy,laplacian))
cv_show('res',res)