OpenCV 入门

ops/2025/3/15 13:33:14/

OpenCV_Python 入门

创建/显示窗口

方法说明参数返回
namedWindow创建一个窗口窗口名称,WINDOW_NORMAL(窗口属性)
resizeWindow设置窗口大小窗口名称,宽,高
imshow显示窗口窗口名称,要显示的图像
destroyAllWindows关闭所有窗口
waitKey延时监听键盘按下0无限 / 单位毫秒键盘按下键的ASCII码
import cv2cv2.namedWindow("img",cv2.WINDOW_NORMAL)
cv2.resizeWindow("img",800,600)while True:cv2.imshow("img",0)key = cv2.waitKey(0)if key & 0xFF == ord('q'):break
cv2.destroyAllWindows()

显示图片

方法说明参数返回
imread读取图片路径,IMREAD_GRAYSCALE(灰)/IMREAD_COLOR(彩)/其它读取到的图片资源
import cv2while True:# cv2.imshow("img",cv2.imread("../1.jpg",cv2.IMREAD_COLOR))cv2.imshow("img",cv2.imread("../1.jpg",cv2.IMREAD_GRAYSCALE))key = cv2.waitKey(0)if key & 0xFF == ord('q'):break
cv2.destroyAllWindows()

保存图片

方法说明参数返回
imwrite保存图片保存的文件名称,图片资源
import  cv2img = cv2.imread("../1.jpg",cv2.IMREAD_COLOR)while True:cv2.imshow("img",img)key = cv2.waitKey(0)if key & 0xFF == ord('q'):breakelif key & 0xFF == ord('s'):cv2.imwrite("11.png",img)
cv2.destroyAllWindows()

视频采集

方法说明参数返回
VideoCapture摄像机0(开启本机摄像头)
VideoCapture.read获取视频帧ret(读取到帧为true),frame(视频帧)
VideoCapture.release释放资源
import cv2cv2.namedWindow("frame",cv2.WINDOW_NORMAL)
cv2.resizeWindow("frame",800,600)
cap = cv2.VideoCapture(0)
while True:ret,frame = cap.read()cv2.imshow('frame',frame)key = cv2.waitKey(1)if key & 0xFF == ord('q'):break
cap.release()
cv2.destroyAllWindows()

读取视频文件

方法说明参数返回
VideoCapture摄像机本机视频文件路径
import cv2cv2.namedWindow("frame",cv2.WINDOW_NORMAL)
cv2.resizeWindow("frame",800,600)
cap = cv2.VideoCapture("../2.mp4")
while True:ret,frame = cap.read()cv2.imshow('frame',frame)key = cv2.waitKey(1)if key & 0xFF == ord('q'):break
cap.release()
cv2.destroyAllWindows()

视频录制

方法说明参数返回
VideoWriter_fourcc编码器*MJPG / 其它
VideoWriter视频输出文件名称,编码器,帧数,(宽,高)
VideoWriter.write视频写入帧
VideoWriter.release释放资源
import cv2
fourcc = cv2.VideoWriter_fourcc(*'MJPG')
vw = cv2.VideoWriter('output.avi', fourcc, 20.0, (640, 480))
cv2.namedWindow("frame",cv2.WINDOW_NORMAL)
cv2.resizeWindow("frame",800,600)
cap = cv2.VideoCapture(0)
while True:ret,frame = cap.read()cv2.imshow('frame',frame)vw.write(frame)key = cv2.waitKey(1)if key & 0xFF == ord('q'):break
cap.release()
vw.release()
cv2.destroyAllWindows()

import cv2
fourcc = cv2.VideoWriter_fourcc(*'MJPG')
vw = cv2.VideoWriter('output.avi', fourcc, 20.0, (640, 480))cv2.namedWindow("frame",cv2.WINDOW_NORMAL)
cv2.resizeWindow("frame",640,480)
cap = cv2.VideoCapture(0)
while cap.isOpened(): # 判断摄像头是否为打开状态ret,frame = cap.read()if not ret: break # 判断帧是否打开cv2.imshow('frame',frame)# 窗口撑开后重新设置cv2.resizeWindow("frame", 640, 480)vw.write(frame)key = cv2.waitKey(1)if key & 0xFF == ord('q'):break
cap.release()
vw.release()
cv2.destroyAllWindows()

鼠标控制

方法说明参数返回
setMouseCallback监听鼠标窗口名称,监听事件,额外参数
import cv2# 鼠标操作    x    y    组合操作    额外参数 
def mouse_callback(event, x, y, flags, param):print(event, x, y, flags, param)
cv2.namedWindow("img",cv2.WINDOW_NORMAL)
cv2.resizeWindow("img",800,600)
cv2.setMouseCallback("img",mouse_callback,"123")
img = cv2.imread("../1.jpg")
while True:cv2.imshow("img",img)key = cv2.waitKey(0)if key & 0xFF == ord('q'):break
cv2.destroyAllWindows()

TrackBar 滑块

方法说明参数返回
createTrackbar创建滑块名称,窗口名称,value 当前值,count 最大值,callback 回调,额外参数
getTrackbarPos获取滑块名称,窗口名称当前值
import  cv2
import numpy as npdef callback():passcv2.namedWindow("img",cv2.WINDOW_NORMAL)cv2.createTrackbar("R","img",0,255,callback)
cv2.createTrackbar("G","img",0,255,callback)
cv2.createTrackbar("B","img",0,255,callback)img = np.zeros((480,640,3),np.uint8)while True:r = cv2.getTrackbarPos("R","img")g = cv2.getTrackbarPos("G","img")b = cv2.getTrackbarPos("B","img")img[:] = [b,g,r]cv2.imshow("img",img)key = cv2.waitKey(10)if key & 0xFF == ord('q'):break
cv2.destroyAllWindows()

色彩

  • opencv 采用 BGR 格式的颜色显示
  • HSV 色相 / 饱和度 / 明度
  • HSL 色相 / 饱和度 / 明度
  • 色彩空间转换
方法说明参数返回
cvtColor色彩转换获取到的图片资源,色彩转换
import cv2def callback():pass
cv2.namedWindow('image',cv2.WINDOW_NORMAL)
cv2.resizeWindow('image',800,600)
img = cv2.imread("../1.jpg",1)
colorspaces = [cv2.COLOR_BGR2RGB,cv2.COLOR_BGR2BGRA,cv2.COLOR_BGR2GRAY,cv2.COLOR_BGR2HSV_FULL,cv2.COLOR_BGR2YUV]
cv2.createTrackbar('curcolor','image',0,len(colorspaces)-1,callback)
while True:v = cv2.getTrackbarPos('curcolor','image')# 颜色空间转换cvt_img = cv2.cvtColor(img,colorspaces[v])cv2.imshow('image',cvt_img)key = cv2.waitKey(10)if key & 0xFF == ord('q'):break
cv2.destroyAllWindows()

Numpy 基本操作

  • 创建矩阵
方法参数说明
array[]
zeros(行,列,通道数),类型值为0
ones(行,列,通道数),类型值为1
full(行,列,通道数),值,类型
identity3三行三列
eye3,5,k=1三行五列, 从第 1-1列开始
import numpy as npa = np.array([1,2,3])
b = np.array([[1,2,3],[4,5,6]])
print(a)
print(b)c = np.zeros((8,8,3),np.uint8)
print(c)d = np.ones((8,8,3),np.uint8)
print(d)e = np.full((8,8,3),10,np.uint8)
print(e)f = np.identity(4)
print(f)y = np.eye(5,7,k=1)
print(y)
  • 赋值 [y,x]
  • 赋值 [y,x,channel]
import numpy as np
import cv2img = np.zeros((640,480,3),np.uint8)
# 检索
print(img[100,100])
count = 0
while count < 100:# 赋值      通道  0,1,2# img[count,100,0] = 255img[count, 100] = [0,0,255]count = count+1
cv2.imshow("img",img)
key = cv2.waitKey(0)
if key & 0xFF == ord('q'):cv2.destroyAllWindows()
  • ROI 获取子矩阵 [y1:y2,x1:x2]
  • [:,:] 整个矩阵 [:]
import numpy as np
import cv2img = np.zeros((640,480,3),np.uint8)
roi = img[100:200,100:200]
img[:] = [255,0,0]
roi[:,:] = [0,0,255]
roi[10:20,10:20] = [255,255,255]
cv2.imshow("img",img)
key = cv2.waitKey(0)
if key & 0xFF == ord('q'):cv2.destroyAllWindows()

Mat

  • 深浅拷贝
import numpy as np
import cv2img = cv2.imread('../1.jpg')
# 浅拷贝
img2 = img
# 深拷贝
img3 = img.copy()
img[10:100,10:100] = [0,0,255]
cv2.imshow('img',img)
cv2.imshow('img2',img2)
cv2.imshow('img3',img3)
cv2.waitKey(0)
  • 属性
import numpy as np
import cv2img = cv2.imread('../1.jpg')
# shape 高,长度,通道
print(img.shape)# 图像占用多大空间
# 高度 * 长度 * 通道数
print(img.size)# 图像中每个元素的位深
print(img.dtype)
  • 分割 / 合并
import numpy as np
import cv2img = np.zeros((480,640,3),np.uint8)# 分割
b,g,r = cv2.split(img)b[10:100,10:100] = 255
g[10:100,10:100] = 255# 合并
img2 = cv2.merge([b,g,r])cv2.imshow("img",img)
cv2.imshow("b",b)
cv2.imshow("g",g)
cv2.imshow("img2",img2)
cv2.waitKey(0)

图形绘制

  • 线
  • 矩形
  • 圆
  • 椭圆
  • 多边形
  • 文本
import numpy as np
import cv2img = np.zeros((480,640,3),np.uint8)# 画线 坐标点 x,y
# 起始点     结束点    bgr     线条宽     线型:平滑
cv2.line(img,(10,20),(300,400),(0,0,255),2,16)# 绘制矩形
# 图像   起始点     结束点    bgr     线条宽 -1为填充
cv2.rectangle(img, (100, 120), (160, 200), (0, 255, 0), 2)
cv2.rectangle(img, (180, 300), (160, 200), (0, 255, 0), -1)# 圆
# 图像   圆点   半径   bgr     线条宽 -1为填充
cv2.circle(img,(300,300), 30, (255,0,0), 5)
cv2.circle(img,(360,360), 30, (255,0,0), -1)# 椭圆
# 图像   圆点   长宽一半  起始角度  终止角度  角度     bgr  线条宽 -1为填充
cv2.ellipse(img,(300,420), (100,30),0,0,360, (255,0,0), 5)
cv2.ellipse(img,(300,420), (100,30),0,0,360, (255,0,0), -1)pts = np.array([(300,10),(150,100),(450,100)],np.int32)
# 多边形
# 图像   点集  是否闭合   bgr  线条宽 
cv2.polylines(img,[pts],True,(0,255,0))
# 多边形填充
# 图像   点集    bgr
cv2.fillPoly(img,[pts],(0,255,0))# 文本
# 图像    文本    起始点    字体    字号    brg
cv2.putText(img,"Hello World!!!",(100,100),cv2.FONT_HERSHEY_PLAIN,3,(255,255,255))cv2.imshow("img",img)
cv2.waitKey(0)

import numpy as np
import cv2# 通过鼠标进行图形的绘制
# 1. l键 画线
# 2. r键 画矩形
# 3. c键 画圆current = 0
startpos = (0,0)
def mouse_callback(event, x, y, flags, param):# print(event, x, y, flags, param)global current,startposif event & cv2.EVENT_LBUTTONDOWN == cv2.EVENT_LBUTTONDOWN:startpos = (x,y)elif event & cv2.EVENT_LBUTTONUP == cv2.EVENT_LBUTTONUP:if current == 0:cv2.line(img,startpos,(x,y),(0,0,255))elif current == 1:cv2.rectangle(img,startpos,(x,y),(0,0,255))elif current == 2:a = (x-startpos[0])b = (y-startpos[1])r = (a**2 + b**2)**0.5cv2.circle(img,startpos,int(r),(0,0,255))else:print("无其它")cv2.namedWindow("img",cv2.WINDOW_NORMAL)cv2.setMouseCallback("img",mouse_callback,"123")img = np.zeros((480,640,3),np.uint8)
while True:cv2.imshow("img",img)key = cv2.waitKey(1)if key & 0xFF == ord('q'):breakelif key  == ord('l'):current =0elif key  == ord('r'):current =1elif key  == ord('c'):current =2cv2.destroyAllWindows()

图像运算

  • 加
  • 减
  • 乘
  • 除
  • 溶合
import numpy as np
import cv2img = cv2.imread('../1.jpg')
img2 = np.ones(img.shape,np.uint8) * 50
# 加    原始图像   相加的图像
img11 = cv2.add(img, img2)
# 减    原始图像   相减的图像
img22 = cv2.subtract(img, img2)
# 乘    原始图像   相乘的图像
img33 = cv2.multiply(img, img2)
# 除    原始图像   相除的图像
img44 = cv2.divide(img, img2)
# 溶合    原始图像 原始图像权重    溶合的图像 溶合的图像权重  静态权重
img55 = cv2.addWeighted(img,0.1, img2,0.9,0)
cv2.imshow('img', img)
cv2.imshow('img11', img11)
cv2.imshow('img22', img22)
cv2.imshow('img33', img33)
cv2.imshow('img44', img44)
cv2.imshow('img55', img55)
cv2.waitKey(0)
  • 非运算
  • 与运算
  • 或运算
  • 异或运算
import numpy as np
import cv2img1 = np.zeros((200,200),np.uint8)
img1[50:150,50:150] = 255
# 非运算
img11 = cv2.bitwise_not(img1)img2 = np.zeros((200,200),np.uint8)
img22 = np.zeros((200,200),np.uint8)
img2[20:120,20:120] = 255
img22[80:150,80:150] = 255
# 与运算
# img222 = cv2.bitwise_and(img2,img22)
# 或运算
# img222 = cv2.bitwise_or(img2,img22)
# 异或运算
img222 = cv2.bitwise_xor(img2,img22)cv2.imshow('img1',img1)
cv2.imshow('img11',img11)
cv2.imshow('img222',img222)
cv2.waitKey(0)

import numpy as np
import cv2img = cv2.imread('../1.jpg')# roi 获取图标在原始图像的位置
roi = img[0:200,0:200]# 图标创建
logo = np.zeros((200,200, 3), np.uint8)
logo[20:120,20:120] = [0,0,255]
logo[80:180,80:180] = [0,255,0]# 创建一个与图标相同的图像
# 使用非运算取相反的图像
# 图标在原始图像的位置与当前获取相反图标进行与运算得到原始图像中显示图标的位置
# logo与原始图像中显示图标相加得到原始图像显示logo的图像
mask = np.zeros((200,200), np.uint8)
mask[20:120,20:120] = 255
mask[80:180,80:180] = 255
m = cv2.bitwise_not(mask)
tmp = cv2.bitwise_and(roi,roi,mask=m)
dst = cv2.add(tmp,logo)# 更改原始图像显示logo位置的图像
img[0:200,0:200] = dstcv2.imshow('logo', dst)
cv2.waitKey(0)

图像变换

  • 缩放
  • 翻转
  • 旋转
import numpy as np
import cv2img = cv2.imread('../1.jpg')# 缩放
# 图像  (长,宽)   fx 长缩放比例  fy 宽缩放比例   interpolation 插值算法
img1 = cv2.resize(img, (224, 224))
img2 = cv2.resize(img, None, fx=0.6, fy=0.6, interpolation=cv2.INTER_AREA)
img3 = cv2.resize(img, None, fx=1.6, fy=1.6, interpolation=cv2.INTER_AREA)cv2.imshow('image', img)
cv2.imshow('image1', img1)
cv2.imshow('image2', img2)
cv2.imshow('image3', img3)
cv2.waitKey(0)

import numpy as np
import cv2img = cv2.imread('../1.jpg')# 翻转
# 图像  0 上下翻转   >0 左右翻转    <0 上下左右翻转
img1 = cv2.flip(img, 0)
img2 = cv2.flip(img, 1)
img3 = cv2.flip(img, -1)cv2.imshow('image', img)
cv2.imshow('image1', img1)
cv2.imshow('image2', img2)
cv2.imshow('image3', img3)
cv2.waitKey(0)

import numpy as np
import cv2img = cv2.imread('../1.jpg')# 旋转
img1 = cv2.rotate(img, cv2.ROTATE_90_CLOCKWISE)cv2.imshow('image', img1)
cv2.waitKey(0)
  • 仿射变换
import numpy as np
import cv2img = cv2.imread('../1.jpg')
h,w,ch = img.shape
print(h,w)
# 平移
# M = np.float32([[1,0,100],[0,1,200]])
# 中心点   旋转角度(逆时针)   缩放
# M = cv2.getRotationMatrix2D((w/2,h/2),15,1.0)
src = np.float32([[200,150],[400,150],[400,500]])
dst = np.float32([[100,200],[300,250],[150,600]])
M = cv2.getAffineTransform(src,dst)
# 修改新图的尺寸dsize(w/2,h/2)
img2 = cv2.warpAffine(img,M,(w,h))cv2.imshow('img',img)
cv2.imshow('img2',img2)
cv2.waitKey(0)
  • 透视变换
import numpy as np
import cv2img = cv2.imread('../1.jpg')src = np.float32([[10,300],[100,300],[20,500],[150,450]])
dst = np.float32([[0,0],[120,0],[0,300],[120,300]])
M = cv2.getPerspectiveTransform(src,dst)
img1 = cv2.warpPerspective(img,M,(300,120))cv2.imshow('img',img)
cv2.imshow('img1',img1)
cv2.waitKey(0)

图像滤波

import numpy as np
import cv2img = cv2.imread('../1.jpg')
kernel = np.ones((5,5),np.uint32) / 25
dst = cv2.filter2D(img,-1,kernel)
cv2.imshow('img',img)
cv2.imshow('dst',dst)
cv2.waitKey(0)
  • 方盒滤波
  • 均值滤波
import numpy as np
import cv2img = cv2.imread('../1.jpg')dst = cv2.blur(img,(5,5))cv2.imshow('img',img)
cv2.imshow('dst',dst)
cv2.waitKey(0)
  • 高斯滤波
import numpy as np
import cv2img = cv2.imread('../1.jpg')dst = cv2.GaussianBlur(img,(5,5),sigmaX=1)cv2.imshow('img',img)
cv2.imshow('dst',dst)
cv2.waitKey(0)
  • 中值滤波
import numpy as np
import cv2img = cv2.imread('../1.jpg')dst = cv2.medianBlur(img,5)cv2.imshow('img',img)
cv2.imshow('dst',dst)
cv2.waitKey(0)
  • 双边滤波
import numpy as np
import cv2img = cv2.imread('../1.jpg')dst = cv2.bilateralFilter(img,5,20,50)cv2.imshow('img',img)
cv2.imshow('dst',dst)
cv2.waitKey(0)
  • 高通滤波
import numpy as np
import cv2img = cv2.imread('./12.png')# 索贝尔算子x方向边缘
dst = cv2.Sobel(img,cv2.CV_64F,1,0,ksize=5)
# 索贝尔算子y方向边缘
dst1 = cv2.Sobel(img,cv2.CV_64F,0,1,ksize=5)
dst2 = cv2.add(dst,dst1)
cv2.imshow('img',img)
cv2.imshow('dst',dst2)
cv2.waitKey(0)

import numpy as np
import cv2img = cv2.imread('./12.png')# 沙尔算子x方向边缘
dst = cv2.Scharr(img,cv2.CV_64F,1,0)
# 沙尔算子y方向边缘
dst1 = cv2.Scharr(img,cv2.CV_64F,0,1)
dst2 = cv2.add(dst,dst1)
cv2.imshow('img',img)
cv2.imshow('dst',dst2)
cv2.waitKey(0)

import numpy as np
import cv2img = cv2.imread('./12.png')# 拉普拉斯算子边缘,需先进行降噪
dst = cv2.Laplacian(img,cv2.CV_64F,ksize=5)cv2.imshow('img',img)
cv2.imshow('dst',dst)
cv2.waitKey(0)

边缘检测

import numpy as np
import cv2img = cv2.imread('../1.jpg')# Canny
dst = cv2.Canny(img,100,200)cv2.imshow('img',img)
cv2.imshow('dst',dst)
cv2.waitKey(0)

形态学图像处理

  • 图像二值化
import numpy as np
import cv2img = cv2.imread('../1.jpg')
img1 = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)ret,img2 = cv2.threshold(img1,180,255,cv2.THRESH_BINARY)cv2.imshow('img',img)
cv2.imshow('img1',img1)
cv2.imshow('img2',img2)cv2.waitKey(0)

import numpy as np
import cv2img = cv2.imread('../1.jpg')
img1 = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)dst = cv2.adaptiveThreshold(img1,255,cv2.ADAPTIVE_THRESH_GAUSSIAN_C,cv2.THRESH_BINARY,3,0)cv2.imshow('img',img)
cv2.imshow('img1',img1)
cv2.imshow('img2',dst)cv2.waitKey(0)
  • 腐蚀
  • 膨胀
import numpy as np
import cv2img = cv2.imread('./img.png')# kernel = np.ones((3,3),np.uint8)
kernel = cv2.getStructuringElement(cv2.MORPH_RECT,(3,3))
# 腐蚀
img1 = cv2.erode(img,kernel,iterations = 1)
# 膨胀
img2 = cv2.dilate(img1,kernel,iterations = 1)cv2.imshow('img',img)
cv2.imshow('img1',img1)
cv2.imshow('img2',img2)cv2.waitKey(0)
  • 开运算
    • 腐蚀 + 膨胀
  • 闭运算
    • 膨胀 + 腐蚀
  • 梯度
  • 顶帽
    • 原图 - 开运算
  • 黑帽
    • 原图 - 闭运算
import numpy as np
import cv2img = cv2.imread('./img_1.png')
img3 = cv2.imread('./img_3.png')
img6 = cv2.imread('./img_2.png')
imgd = cv2.imread('./img_4.png')# kernel = np.ones((3,3),np.uint8)
kernel = cv2.getStructuringElement(cv2.MORPH_RECT,(7,7))
# 开运算
img2 = cv2.morphologyEx(img, cv2.MORPH_OPEN, kernel)
# 闭运算
img33 = cv2.morphologyEx(img3, cv2.MORPH_CLOSE, kernel)
# 梯度
img66 = cv2.morphologyEx(img6, cv2.MORPH_GRADIENT, kernel)
# 顶帽
imgdd = cv2.morphologyEx(imgd, cv2.MORPH_TOPHAT, cv2.getStructuringElement(cv2.MORPH_RECT,(19,19)))
# 黑帽
imgh = cv2.morphologyEx(img3, cv2.MORPH_BLACKHAT, kernel)cv2.imshow('img2',img2)
cv2.imshow('img33',img33)
cv2.imshow('img66',img66)
cv2.imshow('imgdd',imgdd)
cv2.imshow('imgh',imgh)
cv2.waitKey(0)

图像轮廓

import numpy as np
import cv2img = cv2.imread('./img_5.png')# 转为单通道
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
# 二值化
ret,img1 = cv2.threshold(gray,150,255,cv2.THRESH_BINARY)
# 查找轮廓
contours,hierarchy = cv2.findContours(img1,cv2.RETR_TREE,cv2.CHAIN_APPROX_SIMPLE)
# 绘制轮廓
cv2.drawContours(img,contours,-1,(0,0,255),1)
# 计算面积
area = cv2.contourArea(contours[0])
print(area)
# 计算周长
len = cv2.arcLength(contours[0],True)
print(len)cv2.imshow('img', img)
cv2.waitKey(0)

import numpy as np
import cv2img = cv2.imread('./img_6.png')# 转为单通道
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
# 二值化
ret,img1 = cv2.threshold(gray,150,255,cv2.THRESH_BINARY)
# 查找轮廓
contours,hierarchy = cv2.findContours(img1,cv2.RETR_TREE,cv2.CHAIN_APPROX_SIMPLE)
# 绘制轮廓
cv2.drawContours(img,contours,0,(255,0,0),1)
# 多边形逼近
approx = cv2.approxPolyDP(contours[0],20,True)
# 多边形凸包
hull = cv2.convexHull(contours[0])def drawShape(img, approx):i = 0while i < len(approx):if i == len(approx) - 1:x, y = approx[i][0]x1, y1 = approx[0][0]cv2.line(img, (x, y), (x1, y1), (0, 0, 255), 1)else:x, y = approx[i][0]x1, y1 = approx[i + 1][0]cv2.line(img, (x, y), (x1, y1), (0, 0, 255), 1)i = i+1drawShape(img,approx)
drawShape(img,hull)cv2.imshow('img', img)
cv2.waitKey(0)

import numpy as np
import cv2img = cv2.imread('./img_7.png')# 转为单通道
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
# 二值化
ret,img1 = cv2.threshold(gray,150,255,cv2.THRESH_BINARY)
# 查找轮廓
contours,hierarchy = cv2.findContours(img1,cv2.RETR_TREE,cv2.CHAIN_APPROX_SIMPLE)# 最小外接矩形
r = cv2.minAreaRect(contours[1]) 
box = cv2.boxPoints(r)
box = np.int8(box)
cv2.drawContours(img,[box],0,(0,0,255),2)# 最大外接矩形
x,y,w,h = cv2.boundingRect(contours[1])
cv2.rectangle(img,(x,y),(x+w,y+h),(0,255,0),2)cv2.imshow("img",img)
cv2.waitKey(0)

特征点检测

  • 哈里斯角点检测
  • Shi-Tomasi 角点检测
  • SIFT
  • SURF
  • ORB
import numpy as np
import cv2img = cv2.imread('./12.png')
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)# dst = cv2.cornerHarris(gray, 2, 3, 0.04)
# img[dst>0.01*dst.max()] = [0,0,255]# corners = cv2.goodFeaturesToTrack(gray, 1000, 0.01, 10)
# corners = np.int8(corners)
# for i in corners:
#     x,y = i.ravel()
#     cv2.circle(img,(x,y),3,(255,0,0),-1)# sift = cv2.SIFT.create()
# kp,des = sift.detectAndCompute(gray, None)
# cv2.drawKeypoints(gray,kp,img)orb = cv2.ORB.create()
kp,des = orb.detectAndCompute(gray,None)
cv2.drawKeypoints(gray,kp,img)cv2.imshow("img",img)
cv2.waitKey(0)

特征匹配

  • BF
import numpy as np
import cv2img1 = cv2.imread('./img.png')
img2 = cv2.imread('./img_1.png')
gray1 = cv2.cvtColor(img1, cv2.COLOR_BGR2GRAY)
gray2 = cv2.cvtColor(img2, cv2.COLOR_BGR2GRAY)sift = cv2.SIFT.create()
kp1,des1 = sift.detectAndCompute(gray1, None)
kp2,des2 = sift.detectAndCompute(gray2, None)bf = cv2.BFMatcher(cv2.NORM_L1)
matches = bf.match(des1, des2)
img3 = cv2.drawMatches(img1, kp1, img2, kp2, matches, None, flags=2)cv2.imshow("img",img3)
cv2.waitKey(0)
  • FLANN
import numpy as np
import cv2img1 = cv2.imread('./img.png')
img2 = cv2.imread('./img_1.png')
gray1 = cv2.cvtColor(img1, cv2.COLOR_BGR2GRAY)
gray2 = cv2.cvtColor(img2, cv2.COLOR_BGR2GRAY)sift = cv2.SIFT.create()
kp1,des1 = sift.detectAndCompute(gray1, None)
kp2,des2 = sift.detectAndCompute(gray2, None)flann = cv2.FlannBasedMatcher( dict(trees=5,algorithm=1),dict(checks=50))
matches = flann.knnMatch(des1, des2,k=2)good = []
for i,(m,n) in enumerate(matches):if m.distance < 0.7*n.distance:good.append(m)ret = cv2.drawMatchesKnn(img1,kp1,img2,kp2,[good],None)cv2.imshow("img",ret)
cv2.waitKey(0)

图像分割

import numpy as np
import cv2img = cv2.imread("./img.png")
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
ret,thresh = cv2.threshold(gray,0,255,cv2.THRESH_BINARY_INV+cv2.THRESH_OTSU)kernel = np.ones((3,3),np.uint8)
open1 = cv2.morphologyEx(thresh,cv2.MORPH_OPEN,kernel,iterations=2)bg = cv2.dilate(open1,kernel,iterations=2)dist = cv2.distanceTransform(open1,cv2.DIST_L2,5)
re1t,fg = cv2.threshold(dist,0.7*dist.max(),255,cv2.THRESH_BINARY)fg = np.uint8(fg)
unknow = cv2.subtract(bg,fg)ret3,marker = cv2.connectedComponents(fg)
marker = marker + 1
marker[unknow == 255] = 0result = cv2.watershed(img,marker)img[result == -1] = [0,0,255]cv2.imshow("img",img)
cv2.waitKey(0)

import numpy as np
import cv2class App:rect = (0,0,0,0)flag_rect = FalsestartX = 0startY = 0def onmouse(self,event,x,y,flags,param):print("onmouse")if event == cv2.EVENT_LBUTTONDOWN:self.flag_rect = Trueself.startX = xself.startY = yelif event == cv2.EVENT_LBUTTONUP:self.flag_rect = Falsecv2.rectangle(self.img,(self.startX,self.startY),(x,y),(0,0,255),2)self.rect = (min(self.startX,x),min(self.startY,y),abs(self.startX - x),abs(self.startY - y))print("左键抬起")elif event == cv2.EVENT_MOUSEMOVE:if self.flag_rect:self.img = self.img2.copy()cv2.rectangle(self.img,(self.startX,self.startY),(x,y),(255,0,0),2)print("移动")def run(self):print("run")cv2.namedWindow('input')cv2.setMouseCallback('input',self.onmouse)self.img = cv2.imread('./img.png')self.img2 = self.img.copy()self.mask = np.zeros(self.img.shape[:2],np.uint8)self.output = np.zeros(self.img.shape,np.uint8)while 1:cv2.imshow('input',self.img)cv2.imshow('output',self.output)k = cv2.waitKey(100)if k == 27:breakif k == ord('g'):bgmodel = np.zeros((1,65),np.float64)fgmodel = np.zeros((1,65),np.float64)cv2.grabCut(self.img2,self.mask,self.rect,bgmodel,fgmodel,1,cv2.GC_INIT_WITH_RECT)mask2 = np.where((self.mask == 1)|(self.mask == 3),255,0).astype('uint8')self.output = cv2.bitwise_and(self.img2,self.img2,mask=mask2)App().run()
import numpy as np
import cv2img = cv2.imread("./img.png")img2 = cv2.pyrMeanShiftFiltering(img, 20, 30)img3 = cv2.Canny(img2, 150, 300)contours,_ = cv2.findContours(img3, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
cv2.drawContours(img, contours, -1, (0,0,255), 2)cv2.imshow("img", img)
cv2.imshow("img2", img2)
cv2.imshow("img3", img3)cv2.waitKey(0)

import numpy as np
import cv2cap = cv2.VideoCapture('../2.mp4')mog = cv2.createBackgroundSubtractorMOG2()while cap.isOpened():ret, frame = cap.read()fgmask = mog.apply(frame)cv2.imshow('frame', fgmask)if cv2.waitKey(1) & 0xFF == ord('q'):break
cap.release()
cv2.destroyAllWindows()

图像修复

import numpy as np
import cv2img = cv2.imread("./img_1.png")
mask = cv2.imread("./img_2.png",0)
mask = cv2.resize(mask,(img.shape[1],img.shape[0]))
print(img.shape)
print(mask.shape)
dst = cv2.inpaint(img,mask,3,cv2.INPAINT_TELEA)cv2.imshow("img",dst)
cv2.waitKey(0)

哈尔级联法

人脸识别

import numpy as np
import cv2# 创建haar级联器
facer = cv2.CascadeClassifier('D:\SoftWare\py\Lib\site-packages\cv2\data\haarcascade_frontalface_default.xml'
)img = cv2.imread('./img_3.png')
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)faces = facer.detectMultiScale(gray,1.1,5)for (x, y, w, h) in faces:cv2.rectangle(img, (x, y), (x + w, y + h), (255, 0, 0), 2)cv2.imshow('img', img)
cv2.waitKey(0)

import numpy as np
import cv2# 创建haar级联器
facer = cv2.CascadeClassifier('D:\SoftWare\py\Lib\site-packages\cv2\data\haarcascade_frontalface_default.xml'
)img = cv2.VideoCapture(0)while True:ret,frame = img.read()gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)faces = facer.detectMultiScale(gray,1.1,5)for (x, y, w, h) in faces:cv2.rectangle(frame, (x, y), (x + w, y + h), (255, 0, 0), 2)cv2.imshow('img', frame)k = cv2.waitKey(10)if k == 27:break
img.release()
cv2.destroyAllWindows()
  • 眼
import numpy as np
import cv2# 眼
facer = cv2.CascadeClassifier('D:\SoftWare\py\Lib\site-packages\cv2\data\haarcascade_eye.xml'
)img = cv2.imread('./img_3.png')
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)faces = facer.detectMultiScale(gray,1.1,5)for (x, y, w, h) in faces:cv2.rectangle(img, (x, y), (x + w, y + h), (255, 0, 0), 2)cv2.imshow('img', img)
cv2.waitKey(0)

http://www.ppmy.cn/ops/145755.html

相关文章

【虚拟机网络拓扑记录】

【虚拟机网络拓扑记录】

虚拟机网络拓扑记录 虚拟机安装windows到ubuntu的网络拓扑ubuntu到ubuntu里面的虚拟机网络拓扑windows到ubuntu里面的虚拟机网络拓扑 虚拟机安装 本实验宿主机为windos&#xff0c; 安装vmware&#xff0c;虚拟机操作系统使用ubuntu&#xff0c;然后再在ubuntu上面创建新的虚拟…
阅读更多...
Java重要面试名词整理（七）：分库分表

Java重要面试名词整理（七）：分库分表

文章目录 搭建MySQL主从集群GTID同步集群MySQL高可用方案1、MMM2、MHA3、MGR ShardingSphere理解ShardingSphere的核心概念垂直分片与水平分片ShardingSphere实现分库分表的核心概念 常见的分片策略简单INLINE分片算法STANDARD标准分片算法COMPLEX_INLINE复杂分片算法CLASS_BAS…
阅读更多...
WebSocket | 背景 概念 原理 使用 优缺点及适用场景

WebSocket | 背景 概念 原理 使用 优缺点及适用场景

1 背景 在 WebSocket 出现之前&#xff0c;为了实现推送技术&#xff0c;所用的技术都是轮询&#xff0c;轮询是指浏览器每隔一段时间向服务器发出 HTTP 请求&#xff0c;服务器再返回最新的数据给客户端 常见的轮询方式分为轮询与长轮询&#xff0c;它们的区别如下图所示&…
阅读更多...
Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena

Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena

指令微调后的模型不一定在传统Benchmark上取得更好的结果&#xff0c;类似MMLU和HELM。根据人类爱好对齐后的模型&#xff0c;需要新的评测方法。 文章提出了两个主要内容&#xff1a;MT-bench和Chatbot Arena MT-bench是一系列开放式问题&#xff0c;用于评估聊天机器人的多回…
阅读更多...
什么是Spring Boot 应用开发？

什么是Spring Boot 应用开发？

一、引言 在当今的软件开发领域&#xff0c;Java 依然占据着重要的地位&#xff0c;而 Spring Boot 作为 Java 生态系统中极具影响力的框架&#xff0c;极大地简化了企业级应用的开发流程&#xff0c;提升了开发效率和应用的可维护性。它基于 Spring 框架构建&#xff0c;通过约…
阅读更多...
Spring事务回滚

Spring事务回滚

Transactional注解 Transactional作用&#xff1a;就是在当前这个方法执行开始之前来开启事务&#xff0c;方法执行完毕之后提交事务。如果在这个方法执行的过程当中出现了异常&#xff0c;就会进行事务的回滚操作。 Transactional注解&#xff1a;我们一般会在业务层当中来控制…
阅读更多...
电子配件行业的未来之路：产品说明书数字化转型的力量

电子配件行业的未来之路：产品说明书数字化转型的力量

在科技飞速发展的今天&#xff0c;电子配件行业作为科技创新的前沿阵地&#xff0c;正经历着前所未有的变革。从智能手机、平板电脑到智能穿戴设备&#xff0c;各种新型电子配件层出不穷&#xff0c;极大地丰富了人们的生活。然而&#xff0c;随着产品种类的增多和功能的复杂化…
阅读更多...
android——屏幕适配

android——屏幕适配

一、屏幕适配的重要性 在 Android 开发中&#xff0c;屏幕适配是非常关键的。因为 Android 设备具有各种各样的屏幕尺寸、分辨率和像素密度。如果没有进行良好的屏幕适配&#xff0c;应用可能会出现显示不完整、元素拉伸或压缩变形、字体大小不合适等问题&#xff0c;极大地影响…
阅读更多...
最新文章

Copyright @ 2022~2023