#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

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

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

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)

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

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)