#冲刺创作新星#图像金字塔原理与轮廓识别检测 原创

import cv2
import numpy as np
import matplotlib.pyplot as plt
img = cv2.imread("XY.jpg")#读入新一的图像
def cv2_imshow(name,img):#构造展示图像的函数
    cv2.imshow(name,img)
    cv2.waitKey(0)
    cv2.destroyAllWindows()
cv2_imshow("XY",img)#展示原图像
up = cv2.pyrUp(img)#进行高斯放大
cv2_imshow("up",up)#展示图像
down = cv2.pyrDown(up)#进行高斯缩小
cv2_imshow("down",down)#展示图像
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.

import cv2
import numpy as np
import matplotlib.pyplot as plt
img = cv2.imread("XY.jpg")#读入新一的图像
def cv2_imshow(name,img):#构造展示图像的函数
    cv2.imshow(name,img)
    cv2.waitKey(0)
    cv2.destroyAllWindows()
cv2_imshow("XY",img)#展示原图像
down = cv2.pyrDown(img)#高斯缩小
down_up = cv2.pyrUp(down)#高斯放大
down_up = cv2.resize(down_up,(220,165))#调整图像大小
laplacian = img - down_up#拉普拉斯金字塔，相当于在求图像残差
cv2_imshow("laplacian",laplacian)#展示图像
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.

import cv2
import numpy as np
import matplotlib.pyplot as plt
def cv2_imshow(name,img):#构造展示图像的函数
    cv2.imshow(name,img)
    cv2.waitKey(0)
    cv2.destroyAllWindows()
img = cv2.imread("contours.png")#导入原图
cv2_imshow("img",img)#展示原图
img_gray = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)#转化为灰度图
ret,thresh = cv2.threshold(img_gray,120,255,cv2.THRESH_BINARY)#进行二值化处理
cv2_imshow("thresh",thresh)#展示二值化处理后的图像
contours,hierarchy = cv2.findContours(thresh,cv2.RETR_TREE,cv2.CHAIN_APPROX_NONE)#这个版本只返回两个值,轮廓以及轮廓的层级
draw_img = img.copy()#复制一下原图
res = cv2.drawContours(draw_img,contours,-1,(0,0,255))#-1表示画出全部轮廓，可以单独画出某个轮廓
cv2_imshow("res",res)#展示图像
#轮廓特征（面积与周长）
cnt = contours[3]#指定轮廓
area = cv2.contourArea(cnt)#计算面积
perimeter = cv2.arcLength(cnt,True)#计算周长，True表示该图像曲线闭合
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.

import cv2
import numpy as np
import matplotlib.pyplot as plt
def cv2_imshow(name,img):
    cv2.imshow(name,img)
    cv2.waitKey(0)
    cv2.destroyAllWindows()
img = cv2.imread("contours.png")#读入图像
cv2_imshow("img",img)#展示原图像
img_gray = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)#转化为灰度图
ret,thresh = cv2.threshold(img_gray,127,255,cv2.THRESH_BINARY)#转化为二值图像
contours,hierarchy = cv2.findContours(thresh,cv2.RETR_TREE,cv2.CHAIN_APPROX_NONE)#计算轮廓
cnt = contours[0]#选定轮廓，可以自由选择，改为1，2，3等等都可以
x,y,w,h = cv2.boundingRect(cnt)#计算矩形的左上点坐标，图片的高和宽（不一定是正数）
cv2.rectangle(img,(x,y),(x+w,y+h),(0,255,0),2)#给出左上点的和右下点矩形坐标计算出矩形的大小和位置
cv2_imshow("img",img)#展示画好矩形的图像
area = cv2.contourArea(cnt)#计算轮廓面积
rect_area = w*h
extent = float(area)/rect_area#计算轮廓与矩形的面积比
print("轮廓与矩形的面积比为：",extent)
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.

import cv2
import numpy as np
import matplotlib.pyplot as plt
def cv2_imshow(name,img):
    cv2.imshow(name,img)
    cv2.waitKey(0)
    cv2.destroyAllWindows()
img = cv2.imread("lena2.png",0)
template = cv2.imread("face.png",0)
res = cv2.matchTemplate(img,template,1)#img为要检测的图片，template为模板，1表示检测方式
min_val,max_val,min_loc,max_loc = cv2.minMaxLoc(res)#得到最大最小值及其坐标
h,w = template.shape[:2]#得到模板的高和宽用以确定矩形大小
methods = ['cv2.TM_CCOEFF',"cv2.TM_CCOEFF_NORMED","cv2.TM_CCORR","cv2.TM_CCORR_NORMED","cv2.TM_SQDIFF","cv2.TM_SQDIFF_NORMED"]
for i in methods:
    draw_img = img.copy()#复制图像
    method = eval(i)
    res = cv2.matchTemplate(img,template,method)#模板匹配操作
    min_val,max_val,min_loc,max_loc = cv2.minMaxLoc(res)#计算出最大值坐标和最小值坐标
    if i in ["cv2.TM_SQDIFF","cv2.TM_SQDIFF_NORMED"]:#判断是否进行归一化操作，寻找匹配的坐标
        top_left = min_loc#如果进行了归一化操作，则应选择最小值
    else:
        top_left = max_loc#如果未进行归一化操作，则选择最大值
    bottom_right = (top_left[0]+w,top_left[1]+h)#求出右下角矩形坐标
    cv2.rectangle(draw_img,top_left,bottom_right,255,2)#计算矩形的左上点坐标，图片的高和宽（不一定是正数）
    plt.subplot(121),plt.imshow(res,cmap = 'gray')
    plt.xticks([]),plt.yticks([])#隐藏坐标轴
    plt.subplot(122),plt.imshow(draw_img,cmap = 'gray')
    plt.xticks([]),plt.yticks([])#隐藏坐标轴
    plt.suptitle(i)#显示标题
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.

import cv2
import numpy as np
import matplotlib.pyplot as plt
def cv2_imshow(name,img):
    cv2.imshow(name,img)
    cv2.waitKey(0)
    cv2.destroyAllWindows()
img_rgb = cv2.imread("ku_2.png")
img_gray = cv2.imread("ku_2.png",0)#读入灰度图
cv2_imshow("gray",img_gray)
template = cv2.imread("god_2.png",0)
res = cv2.matchTemplate(img_gray,template,cv2.TM_CCOEFF_NORMED)
h,w = template.shape[:2]
threshold = 0.5
#取匹配程度大于50%的坐标
loc = np.where(res>threshold)
for pt in zip(*loc[::-1]):
    bottom_right = (pt[0]+w,pt[1]+h)
    cv2.rectangle(img_rgb,pt,bottom_right,(0,225,0),1)
cv2_imshow("img_rgb",img_rgb)
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.