图像的变换，平移旋转放缩(双线性，最近邻域)镜像 python

土豆又来水文章了
今天是图像的变换大礼包，用的是jupyter lab anaconda

干啥啥不行，导库第一名

import cv2
import numpy as np
import random
from matplotlib import pyplot as plt
from numpy import *

这个是最近领域法放缩的函数块

def NN_interpolation(img,dH,dW):
    scrH=img.shape[0]
    scrW=img.shape[1]
    retimg=np.zeros((dH,dW,3),dtype=np.uint8)
    for i in range(dH-1):
        for j in range(dW-1):
            scrx=round(i*(scrH/dH))
            scry=round(j*(scrW/dW))
            retimg[i,j]=img[scrx,scry]
    return retimg

这个是双线性插值法的函数块

def BiLinear_interpolation(img,dH,dW):
    scrH=img.shape[0]
    scrW=img.shape[1]
    img=np.pad(img,((0,1),(0,1),(0,0)),'constant')
    retimg=np.zeros((dH,dW,3),dtype=np.uint8)
    for i in range(dH):
        for j in range(dW):
            scrx=(i+1)*(scrH/dH)-1
            scry=(j+1)*(scrW/dW)-1
            x=math.floor(scrx)
            y=math.floor(scry)
            u=scrx-x
            v=scry-y
            retimg[i,j]=(1-u)*(1-v)*img[x,y]+u*(1-v)*img[x+1,y]+(1-u)*v*img[x,y+1]+u*v*img[x+1,y+1]
    return retimg

这个是平移的函数块

def Move_interpolation(img,mH,mW):
    scrH=img.shape[0]
    scrW=img.shape[1]
    retimg=np.zeros((scrH+mH,scrH+mW,3),dtype=np.uint8)
    for i in range(scrH+mH):
        for j in range(scrH+mW):
            if i-mH>=0 and j-mW>=0:
                scrx=i-mH
                scry=j-mW
                retimg[i,j]=img[scrx,scry]
            else:
                retimg[i,j]=255
    return retimg

这个是令人儒雅随后的旋转函数块

def RotateImage(img,angle):
    scrH=img.shape[0]
    scrW=img.shape[1]
    angle=angle*math.pi/180.0
    cos_a=math.cos(angle)
    sin_a=math.sin(angle)
    X_max1=2*math.ceil(abs(0.5*scrH*cos_a+0.5*scrW*sin_a))
    X_max2=2*math.ceil(abs(0.5*scrH*cos_a-0.5*scrW*sin_a))
    Y_max1=2*math.ceil(abs(0.5*scrH*sin_a+0.5*scrW*cos_a))
    Y_max2=2*math.ceil(abs(0.5*scrH*sin_a-0.5*scrW*cos_a))
    H=int(max(Y_max1,Y_max2))
    W=int(max(X_max1,X_max2))
    retimg=np.zeros((H,W,3),dtype=np.uint8)
    for i in range(H):
        for j in range(W):
                x = int((i-0.5*W)*cos_a-(j-0.5*W)*sin_a+0.5*scrW)
                y = int((i-0.5*H)*sin_a+(j-0.5*W)*cos_a+0.5*scrH)
                if x>=0 and x<scrW and y>=0 and y<scrH:
                    retimg[i,j]=img[x,y]
    return retimg

这个是镜像的函数块，直接对x轴y轴一起镜像

def Mirror(img):
    scrH=img.shape[0]
    scrW=img.shape[1]
    retimg=np.zeros((scrH,scrW,3),dtype=np.uint8)
    for i in range(scrH):
        for j in range(scrW):
            retimg[scrH-i-1,scrW-j-1] = img[i, j]
    return retimg

调用这些函数

img = cv2.imread("newpic.bmp")
out1 = NN_interpolation(img, int(img.shape[0]*2),int(img.shape[1]*2))
plt.title("yuantu")
plt.imshow(img)

plt.title("NNIP")
plt.imshow(out1)
plt.show()

out2 = BiLinear_interpolation(img, int(img.shape[0]*2),int(img.shape[1]*2))
plt.title("BiLinear_interpolation")
plt.imshow(out2)
plt.show()

out3=Move_interpolation(img, 50,50)
plt.title("Move_interpolation")
plt.imshow(out3)
plt.show()

out4 = RotateImage(img,30)
plt.title("RotateImage")
plt.imshow(out4)
plt.show()

out5 = Mirror(img)
plt.title("Mirror")
plt.imshow(out5)
plt.show()

这样就完成了，用jupyter编译的 我还放了我的jupyter文件，里面有已经运行好的结果
添加链接描述
用这个图片跑的，正方形的 1024*1024 像素