1、颜色转换
以红色图像为例。在图像中提取红色部分颜色。
import cv2
import numpy as np
import os
os.chdir(r'C:\Users\Eddie\Desktop')
img = cv2.imread('test111.jpg')
#各种滤波
blur = cv2.blur(img,(5,5))
blur0 = cv2.medianBlur(blur,5)
blur1 = cv2.GaussianBlur(blur0,(5,5),0)
blur2 = cv2.bilateralFilter(blur1,9,75,75)
# 颜色转换，RGB->HSV
hsv = cv2.cvtColor(blur2,cv2.COLOR_BGR2HSV)
# 在HSV空间的颜色范围，参考下表中的颜色区间
lower_red = np.array([156,20,70])
upper_red = np.array([180,255,255])
mask = cv2.inRange(hsv,lower_red,upper_red)
# 与原图与操作，只剩下希望留下的颜色
res = cv2.bitwise_and(img,img,mask = mask)
cv2.imshow('blur',res)
cv2.waitKey(0)
cv2.destroyAllWindows()
image.png
image.png
2. 二值化处理
1） 二值化处理
# 手动二值化处理
# 设置阈值大小 threshold
thresh = 125
# 设置超过阈值像素值的最大值
maxval = 255
# THRESH_BINARY：超过阈值为maxval，否则为0
# THRESH_BINARY_INV：超过阈值为0，否则为maxval（相当于上个参数取反）
# THRESH_TRUNC：超过阈值为thresh，低于阈值灰度值不变
# THRESH_TOZERO：超过阈值灰度值不变，否则为0
# THRESH_TOZERO_INV：超过阈值为0，低于阈值灰度值不变
ret, thresh1 = cv2.threshold(img, thresh, maxval, cv2.THRESH_BINARY)
ret, thresh2 = cv2.threshold(img, thresh, maxval, cv2.THRESH_BINARY_INV)
ret, thresh3 = cv2.threshold(img, thresh, maxval, cv2.THRESH_TRUNC)
ret, thresh4 = cv2.threshold(img, thresh, maxval, cv2.THRESH_TOZERO)
ret, thresh5 = cv2.threshold(img, thresh, maxval, cv2.THRESH_TOZERO_INV)
titles = ['original image', 'Binary', 'binary-inv', 'trunc', 'tozero', 'tozero-inv']
images = [img, thresh1, thresh2, thresh3, thresh4, thresh5]
for i in range(6):
plt.subplot(2, 3, i+1)
plt.imshow(images[i], 'gray')
plt.title(titles[i])
plt.xticks([])
plt.yticks([])
plt.show()
2） 自动二值化处理
img2 = cv2.imread('002.tif', 0)
# rows, cols = img2.shape
# Otsu阈值
ret1, thresh1 = cv2.threshold(img, 0, 255, cv2.THRESH_OTSU)
# Otsu可与其他二值化方式一起使用
ret2, thresh2 = cv2.threshold(blur, 0, 255, cv2.THRESH_OTSU + cv2.THRESH_BINARY)
# # Triangle阈值-很少用到
# ret3, thresh3 = cv2.threshold(img2, 0, 255, cv2.THRESH_TRIANGLE)
# ret4, thresh4 = cv2.threshold(blur, 0, 255, cv2.THRESH_TRIANGLE)
titles = ['original image', "Otsu's", "Otsu's+Binary"] # "Triangle", "Triangle-Otsu's"
images = [img, thresh1, thresh2] # thresh3, thresh4
for i in range(3):
plt.subplot(1, 3, i+1)
plt.imshow(images[i], 'gray')
plt.title(titles[i])
plt.xticks([])
plt.yticks([])
plt.show()
3） 自适应二值化
# maxval为像素值的最大值默认设为255
# blockSize为计算像素阈值的像素邻域的大小：3、5、7。。。。、99
# C为从平均值或加权平均值中减去常数，一般设置为0-50
blockSize = 25
C = 5
# 高斯BINARY
thresh1 = cv2.adaptiveThreshold(img2, maxval, cv2.ADAPTIVE_THRESH_GAUSSIAN_C, cv2.THRESH_BINARY, blockSize, C)
# 高斯BINARY_INV
thresh2 = cv2.adaptiveThreshold(img2, maxval, cv2.ADAPTIVE_THRESH_GAUSSIAN_C, cv2.THRESH_BINARY_INV, blockSize, C)
# 平均BINARY
thresh3 = cv2.adaptiveThreshold(img2, maxval, cv2.ADAPTIVE_THRESH_MEAN_C, cv2.THRESH_BINARY, blockSize, C)
# 平均BINARY_INV
thresh4 = cv2.adaptiveThreshold(img2, maxval, cv2.ADAPTIVE_THRESH_MEAN_C, cv2.THRESH_BINARY_INV, blockSize, C)
titles = ['original image', "gaussian binary", "gaussian binary_inv", "mean binary", "mean binary_inv"]
images = [img2, thresh1, thresh2, thresh3, thresh4]
for i in range(5):
plt.subplot(2, 3, i+1)
plt.imshow(images[i], 'gray')
plt.title(titles[i])
plt.xticks([])
plt.yticks([])
plt.show()
3. 掩码处理
如果摄像头固定，而监控目标位置也相对固定，则可以利用掩码预处理图像，使得计算机只处理某个区域的图像。
mask = np.zeros(np.shape(img),dtype=np.uint8)
mask[200:770,190:260]=255 # 原点在左上角，x向下，y向右
img= cv2.add(img,np.zeros(np.shape(img),dtype=np.uint8),mask=mask)
cv2.imshow('mask',img)
cv2.waitKey(0)
4. 直线检测之霍夫变换
前提：边缘检测完成。
cv2.HoughLines()
dst: 输出图像. 它应该是个灰度图 (但事实上是个二值化图)
lines: 储存着检测到的直线的参数对 (r,\theta) 的容器
rho : 参数极径 r 以像素值为单位的分辨率. 我们使用 1 像素.
theta: 参数极角 \theta 以弧度为单位的分辨率. 我们使用 1度 (即CV_PI/180)
threshold: 设置阈值： 一条直线所需最少的的曲线交点
srn and stn: 参数默认为0
cv2.HoughLinesP(dst, lines, 1, CV_PI/180, 50, 50, 10 )
dst: 输出图像. 它应该是个灰度图 (但事实上是个二值化图)
lines: 储存着检测到的直线的参数对 (x_{start}, y_{start}, x_{end}, y_{end}) 的容器
rho : 参数极径 r 以像素值为单位的分辨率. 我们使用 1 像素.
theta: 参数极角 \theta 以弧度为单位的分辨率. 我们使用 1度 (即CV_PI/180)
threshold: 设置阈值： 一条直线所需最少的的曲线交点。超过设定阈值才被检测出线段，值越大，基本上意味着检出的线段越长，检出的线段个数越少。
minLinLength: 能组成一条直线的最少点的数量. 点数量不足的直线将被抛弃.
maxLineGap: 能被认为在一条直线上的两点的最大距离。
5. 边缘检测
Canny
edge = cv2.Canny(image, threshold1, threshold2[, edges[, apertureSize[, L2gradient ]]])
必要参数：
- 第一个参数是需要处理的原图像，该图像必须为单通道的灰度图；
- 第二个参数是阈值1；
- 第三个参数是阈值2。
Canny检测只能针对灰度图，使用代码如下：
import cv2
import numpy as np
img = cv2.imread("D:/image_gray.jpg", 0)
img = cv2.GaussianBlur(img,(3,3),0)
canny = cv2.Canny(img, 50, 150)
cv2.imshow('Canny', canny)
cv2.waitKey(0)
cv2.destroyAllWindows()
在github上有一个可以在运行时调整阈值大小的程序。其代码如下：
import cv2
import numpy as np
def CannyThreshold(lowThreshold):
detected_edges = cv2.GaussianBlur(gray,(3,3),0)
detected_edges = cv2.Canny(detected_edges,lowThreshold,lowThreshold*ratio,apertureSize = kernel_size)
dst = cv2.bitwise_and(img,img,mask = detected_edges) # just add some colours to edges from original image.
cv2.imshow('canny demo',dst)
lowThreshold = 0
max_lowThreshold = 100
ratio = 3
kernel_size = 3
img = cv2.imread('D:/lion.jpg')
gray = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)
cv2.namedWindow('canny demo')
cv2.createTrackbar('Min threshold','canny demo',lowThreshold, max_lowThreshold, CannyThreshold)
CannyThreshold(0) # initialization
if cv2.waitKey(0) == 27:
cv2.destroyAllWindows()
6. 直线检测（FLD）
Fast Line Detector: opencv-contrib-python模块才行，不是opencv-python模块。
def FLD(image):
# Create default Fast Line Detector class
fld = cv2.ximgproc.createFastLineDetector()
# Get line vectors from the image
lines = fld.detect(image)
# Draw lines on the image
line_on_image = fld.drawSegments(image, lines)
# Plot
plt.imshow(line_on_image, interpolation='nearest', aspect='auto')
plt.show()
return line_on_image
其中 lines = fld.detect(image)返回的是各个线段的顶点坐标。四个一组。很多直线被分成了很多线段，需要重新整合一下才好。
cv2.ximgproc.createFastLineDetector()可以带参数，参与意义如下：
retval = cv.ximgproc.createFastLineDetector( [, _length_threshold[, _distance_threshold[, _canny_th1[, _canny_th2[, _canny_aperture_size[, _do_merge]]]]]] )
Parameters
_length_threshold 10 - Segment shorter than this will be discarded
_distance_threshold 1.41421356 - A point placed from a hypothesis line segment farther than this will be regarded as an outlier
_canny_th1 50 - First threshold for hysteresis procedure in Canny()
_canny_th2 50 - Second threshold for hysteresis procedure in Canny()
_canny_aperture_size 3 - Aperturesize for the sobel operator in Canny()
_do_merge false - If true, incremental merging of segments will be perfomred
功能：找出检测结果中最长的一根直线，绘制该直线及其两个端点，并且在该直线一个端点处绘制箭头。
import cv2
import numpy as np
from scipy.spatial import distance as dist
img = cv2.imread('sDQLM.png')
gray = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)
#Create default Fast Line Detector (FSD)
fld = cv2.ximgproc.createFastLineDetector()
#Detect lines in the image
lines = fld.detect(gray)
dMax = 0
bx_Max = 0
by_Max = 0
ex_Max = 0
ey_Max = 0
for L in lines:
bx,by,ex,ey = L[0]
# compute the Euclidean distance between the two points,
D = dist.euclidean((bx, by), (ex, ey))
if D > dMax:
dMax = D
bx_Max = bx
by_Max = by
ex_Max = ex
ey_Max = ey
lineMax = np.array([[[bx_Max, by_Max, ex_Max,ey_Max]]])
#Draw detected lines in the image
drawn_img = fld.drawSegments(gray,lineMax,True)
cv2.circle(drawn_img, (bx_Max, by_Max), 1, (255,0,0), 2)#line begin
cv2.circle(drawn_img, (ex_Max, ey_Max), 1, (0,255,0), 2)#line end
cv2.imshow("FLD", drawn_img)
cv2.waitKey(0)
7. 录像
import cv2
import numpy as np
import time, datetime
import sys
def camera_record(fnamesave):
"""
函数名：cv2.VideoCapture()
功 能：通过摄像头捕获实时图像数据
返回值：有
参数一：摄像头代号，0为默认摄像头，笔记本内建摄像头一般为 0
或者填写视频名称直接加载本地视频文件
"""
url = "rtsp://admin:zxcvbnm123@192.168.121.8/Streaming/Channels/2" # 可以使用通道1或者2
cap = cv2.VideoCapture(url) # 创建一个 VideoCapture 对象,打开摄像头
time_begin = time.time()
"""
函数名：cap.set( propId , value )
功 能：设置视频参数。设置视频的宽高值和摄像头有关
使用笔记本内置摄像头时只能设置为 1280*720 以及 640*480，哪怕设置值不同时都会自动校正过来，并且仍然返回 True
返回值：布尔值
参数一：需要设置的视频参数
参数二：设置的参数值
"""
# cap.set(3, 480)
# cap.set(4, 320)
flag = 1; # 设置一个标志，用来输出视频信息
"""
函数名：cv2.isOpened()
功 能：返回一个布尔值（ True / False ），检查是否初始化成功，成功返回 True
返回值：布尔值
"""
# fourcc = cv2.VideoWriter_fourcc(*'XVID')#视频存储的格式
# fourcc = cv2.VideoWriter_fourcc(*'MJPG')
fourcc = cv2.VideoWriter_fourcc('m', 'p', '4', 'v')
fps = cap.get(cv2.CAP_PROP_FPS) # 帧率
# 视频的宽高
size = (int(cap.get(cv2.CAP_PROP_FRAME_WIDTH)), \
int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT)))
# print(fnamesave)
out = cv2.VideoWriter(fnamesave, fourcc, fps, size) # 视频存储
while (out.isOpened()): # 循环读取每一帧
"""
函数名：cap.read()
功 能：返回两个值
先返回一个布尔值，如果视频读取正确，则为 True，如果错误，则为 False，也可用来判断是否到视频末尾
再返回一个值，为每一帧的图像，该值是一个三维矩阵
通用接收方法为：
ret,frame = cap.read();
这样 ret 存储布尔值，frame 存储图像
若使用一个变量来接收两个值，如
frame = cap.read()
则 frame 为一个元组，原来使用 frame 处需更改为 frame[1]
返回值：R1：布尔值
R2：图像的三维矩阵
"""
if time.time()-time_begin > 3600: # 3600秒，即一个小时
return 0
ret_flag, Vshow = cap.read()
# gray = cv2.cvtColor(Vshow,cv2.COLOR_BGR2GRAY)
# cv2.imshow("Gray",gray)
if ret_flag is False: # ret为false则关闭
exit()
cv2.namedWindow('video', cv2.WINDOW_AUTOSIZE) # 创建一个名为video的窗口
cv2.imshow("video", Vshow) # 窗口显示，显示名为 Capture_Test
out.write(Vshow) # 将捕捉到的图像存储
k = cv2.waitKey(1) & 0xFF # 每帧数据延时 1ms，延时不能为 0，否则读取的结果会是静态帧
if k == ord('s'): # 若检测到按键 ‘s’，打印字符串
"""
函数名：cap.get( propId )
功 能：查询视频的参数
返回值：无
参数一：查询的视频参数，其中部分值可以使用 cap.set() 进行修改
"""
print(cap.get(3))
print(cap.get(4))
elif k == ord('q'): # 若检测到按键 ‘q’，退出
sys.exit(0)
break
cap.release() # 释放摄像头
out.release()
cv2.destroyAllWindows() # 删除建立的全部窗口
if __name__ == "__main__":
while True:
fname=datetime.datetime.now().strftime('outvideo%Y-%m-%d-%H') # 一个小时换一次文件名
fdir = r'D:/'
fnamesave = fdir+fname+r'.mp4'
camera_record(fnamesave)
8. 摄像头设置cap.set()
import cv2
#选择摄像头号，一般从 0 开始
cap = cv2.VideoCapture(0)
#先设置参数，然后读取参数
for i in range(47):
print("No.={} parameter={}".format(i,cap.get(i)))
while True:
ret, img = cap.read()
cv2.imshow("input", img)
# 按 ESC 键退出
key = cv2.waitKey(10)
if key == 27:
break
cv2.destroyAllWindows()
cv2.VideoCapture(0).release()
运行结果可以看到，那些不是-1的，就是可以设置。关于序号的说明如下：
0. CV_CAP_PROP_POS_MSEC Current position of the video file in milliseconds.
1. CV_CAP_PROP_POS_FRAMES 0-based index of the frame to be decoded/captured next.
2. CV_CAP_PROP_POS_AVI_RATIO Relative position of the video file
3. CV_CAP_PROP_FRAME_WIDTH Width of the frames in the video stream.
4. CV_CAP_PROP_FRAME_HEIGHT Height of the frames in the video stream.
5. CV_CAP_PROP_FPS Frame rate.
6. CV_CAP_PROP_FOURCC 4-character code of codec.
7. CV_CAP_PROP_FRAME_COUNT Number of frames in the video file.
8. CV_CAP_PROP_FORMAT Format of the Mat objects returned by retrieve() .
9. CV_CAP_PROP_MODE Backend-specific value indicating the current capture mode.
10. CV_CAP_PROP_BRIGHTNESS Brightness of the image (only for cameras).
11. CV_CAP_PROP_CONTRAST Contrast of the image (only for cameras).
12. CV_CAP_PROP_SATURATION Saturation of the image (only for cameras).
13. CV_CAP_PROP_HUE Hue of the image (only for cameras).
14. CV_CAP_PROP_GAIN Gain of the image (only for cameras).
15. CV_CAP_PROP_EXPOSURE Exposure (only for cameras).
16. CV_CAP_PROP_CONVERT_RGB Boolean flags indicating whether images should be converted to RGB.
17. CV_CAP_PROP_WHITE_BALANCE Currently unsupported
18. CV_CAP_PROP_RECTIFICATION Rectification flag for stereo cameras (note: only supported by DC1394 v 2.x backend currently)
9. 三维显示灰度图
Y = np.arange(0, np.shape(img_gray)[0], 1)
X = np.arange(0, np.shape(img_gray)[1], 1)
X, Y = np.meshgrid(X, Y)
fig = plt.figure()
ax = plt.axes(projection="3d")
ax.plot_surface(X, Y, img_gray, cmap='hot') #'gist_rainbow') # cmap='hot'
plt.show()
说明：输入灰度图（img_gray），X, Y是图像矩阵的两个维度。灰度值是第三维。可以看到灰度和人眼感知之间的关系。
网友评论