67 lines
1.9 KiB
Python
67 lines
1.9 KiB
Python
# -*- coding: utf-8 -*-
|
||
# @Time : 2022-7-26 0026 19:09
|
||
# @Author : Qing
|
||
# @Email : derighoid@gmail.com
|
||
# @File : Blue_noise_sampling.py
|
||
# @Software: PyCharm
|
||
from pathlib import Path
|
||
|
||
import matplotlib
|
||
import matplotlib.pyplot as plt
|
||
import numpy as np
|
||
from PIL import Image
|
||
|
||
'''
|
||
最近邻插值法:
|
||
在放大图像时,多出来的像素点由最近邻的像素点构成
|
||
计算新图形(放大后或缩小后)的坐标点像素值对应于原图像中哪一个像素点填充的。
|
||
srcX=newX*(srcW/newW)
|
||
srcY=newY*(srcH/newH)
|
||
src是原图,dst是新图,原来的图像宽度/高度除以新图像的宽度/高度可以得到缩放比例,
|
||
假如是缩小图片括号内的数字小于1,放大则大于1,相当于系数,再乘以新图片的宽度/高度,就实现了缩放。
|
||
'''
|
||
|
||
def blueNoiseSampl(img, newH, newW):
|
||
'''
|
||
|
||
:param img: 图片
|
||
:param newH: 新图的高
|
||
:param newW: 新图的宽
|
||
:return: 新图
|
||
'''
|
||
scrH, scrW, t = img.shape # 原图的长宽
|
||
retimg = np.zeros((newH, newW, 3), dtype=np.uint8) # 生成 newH* newW *3 的零矩阵
|
||
for i in range(newH - 1):
|
||
for j in range(newW - 1):
|
||
scrx = round(i * (scrH / newH)) # round对其四舍五入
|
||
scry = round(j * (scrW / newW))
|
||
retimg[i, j] = img[scrx, scry] # new image
|
||
return retimg
|
||
|
||
|
||
# 图片展示函数
|
||
def showImage(picPath):
|
||
'''
|
||
|
||
:param picPath: 图片地址
|
||
:return: 样图
|
||
'''
|
||
# 获取图片矩阵
|
||
image = np.array(Image.open(picPath))
|
||
# 设置画布
|
||
plt.figure(figsize=(16, 8))
|
||
# 合并
|
||
plt.subplot(121)
|
||
plt.imshow(image)
|
||
# 调用采样函数
|
||
image1 = blueNoiseSampl(image, image.shape[0] * 2, image.shape[1] * 2)
|
||
# 图片保存
|
||
Image.fromarray(np.uint8(image1)).save("./data/picture13.png")
|
||
plt.subplot(122)
|
||
plt.imshow(image1)
|
||
plt.show()
|
||
return image1
|
||
|
||
|
||
|