pos97em56
/
Image_Process
1
0
Fork 1
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Update image_smooth.py

Browse Source
main
pos97em56 5 months ago
parent 6c1bb98f70
commit 4dc8b82e85
1 changed files with 325 additions and 335 deletions
Show Stats Download Patch File Download Diff File

660
basic/image_smooth.py
Unescape View File

@ -1,336 +1,326 @@
import tkinter as tk import tkinter as tk
from tkinter import filedialog, messagebox from tkinter import filedialog, messagebox
from tkinter import Toplevel from tkinter import Toplevel
from PIL import Image, ImageTk from PIL import Image, ImageTk
import numpy as np import numpy as np
import cv2 import cv2
import os import os
img_path = "" # 全局变量，用于存储图像路径 img_path = "" # 全局变量，用于存储图像路径
src = None # 全局变量，用于存储已选择的图像 src = None # 全局变量，用于存储已选择的图像
img_label = None # 全局变量，用于存储显示选择的图片的标签 img_label = None # 全局变量，用于存储显示选择的图片的标签
edge = None edge = None
FreqsmoWin = None FreqsmoWin = None
AirsmoWin = None AirsmoWin = None
def select_image(root): def select_image(root):
global img_path, src, img_label, edge global img_path, src, img_label, edge
img_path = filedialog.askopenfilename(filetypes=[("Image files", "*.jpg;*.png;*.jpeg;*.bmp")]) img_path = filedialog.askopenfilename(filetypes=[("Image files", "*.jpg;*.png;*.jpeg;*.bmp")])
if img_path: if img_path:
# 确保路径中的反斜杠正确处理，并使用 UTF-8 编码处理中文路径 # 确保路径中的反斜杠正确处理，并使用 UTF-8 编码处理中文路径
img_path_fixed = os.path.normpath(img_path) img_path_fixed = os.path.normpath(img_path)
# 图像输入 # 图像输入
src_temp = cv2.imdecode(np.fromfile(img_path_fixed, dtype=np.uint8), cv2.IMREAD_UNCHANGED) src_temp = cv2.imdecode(np.fromfile(img_path_fixed, dtype=np.uint8), cv2.IMREAD_UNCHANGED)
if src_temp is None: if src_temp is None:
messagebox.showerror("错误", "无法读取图片，请选择有效的图片路径") messagebox.showerror("错误", "无法读取图片，请选择有效的图片路径")
return return
src = cv2.cvtColor(src_temp, cv2.COLOR_BGR2RGB) src = cv2.cvtColor(src_temp, cv2.COLOR_BGR2RGB)
# 检查 img_label 是否存在且有效 # 检查 img_label 是否存在且有效
if img_label is None or not img_label.winfo_exists(): if img_label is None or not img_label.winfo_exists():
img_label = tk.Label(root) img_label = tk.Label(root)
img_label.pack(side=tk.TOP, pady=10) img_label.pack(side=tk.TOP, pady=10)
img = Image.open(img_path) img = Image.open(img_path)
img.thumbnail((160, 160)) img.thumbnail((160, 160))
img_tk = ImageTk.PhotoImage(img) img_tk = ImageTk.PhotoImage(img)
img_label.configure(image=img_tk) img_label.configure(image=img_tk)
img_label.image = img_tk img_label.image = img_tk
# 定义 edge 变量为 PIL.Image 对象，以便稍后保存 # 定义 edge 变量为 PIL.Image 对象，以便稍后保存
edge = Image.fromarray(src) edge = Image.fromarray(src)
else: else:
messagebox.showerror("错误", "没有选择图片路径") messagebox.showerror("错误", "没有选择图片路径")
def show_selected_image(root): def changeSize(event, img, LabelPic):
global img_label img_aspect = img.shape[1] / img.shape[0]
img_label = tk.Label(root) new_aspect = event.width / event.height
img_label.pack(side=tk.TOP, pady=10)
img = Image.open(img_path) if new_aspect > img_aspect:
img.thumbnail((160, 160)) new_width = int(event.height * img_aspect)
img_tk = ImageTk.PhotoImage(img) new_height = event.height
img_label.configure(image=img_tk) else:
img_label.image = img_tk new_width = event.width
new_height = int(event.width / img_aspect)
def changeSize(event, img, LabelPic):
img_aspect = img.shape[1] / img.shape[0] resized_image = cv2.resize(img, (new_width, new_height))
new_aspect = event.width / event.height image1 = ImageTk.PhotoImage(Image.fromarray(resized_image))
LabelPic.image = image1
if new_aspect > img_aspect: LabelPic['image'] = image1
new_width = int(event.height * img_aspect)
new_height = event.height def savefile():
else: global edge
new_width = event.width
new_height = int(event.width / img_aspect) filename = filedialog.asksaveasfilename(defaultextension=".jpg", filetypes=[("JPEG files", "*.jpg"), ("PNG files", "*.png"), ("BMP files", "*.bmp")])
if not filename:
resized_image = cv2.resize(img, (new_width, new_height)) return
image1 = ImageTk.PhotoImage(Image.fromarray(resized_image)) # 确保 edge 变量已定义
LabelPic.image = image1 if edge is not None:
LabelPic['image'] = image1 try:
edge.save(filename)
def savefile(): messagebox.showinfo("保存成功", "图片保存成功！")
global edge except Exception as e:
messagebox.showerror("保存失败", f"无法保存图片: {e}")
filename = filedialog.asksaveasfilename(defaultextension=".jpg", filetypes=[("JPEG files", "*.jpg"), ("PNG files", "*.png"), ("BMP files", "*.bmp")]) else:
if not filename: messagebox.showerror("保存失败", "没有图像可保存")
return
# 确保 edge 变量已定义 #频域平滑
if edge is not None: def freq_smo(root):
try: global src, FreqsmoWin, edge
edge.save(filename)
messagebox.showinfo("保存成功", "图片保存成功！") # 判断是否已经选取图片
except Exception as e: if src is None:
messagebox.showerror("保存失败", f"无法保存图片: {e}") messagebox.showerror("错误", "没有选择图片！")
else: return
messagebox.showerror("保存失败", "没有图像可保存")
# 理想低通滤波器
#频域平滑 def Ideal_LowPassFilter(rows, cols, crow, ccol, D0=20):
def freq_smo(root): # 创建一个与输入图像大小相同的空白图像
global src, FreqsmoWin, edge Ideal_LowPass = np.zeros((rows, cols), dtype=np.uint8)
# 创建理想低通滤波器
# 判断是否已经选取图片 for i in range(rows):
if src is None: for j in range(cols):
messagebox.showerror("错误", "没有选择图片！") x = i - crow
return y = j - ccol
D = np.sqrt(x**2 + y**2)
# 理想低通滤波器 if D <= D0:
def Ideal_LowPassFilter(rows, cols, crow, ccol, D0=20): Ideal_LowPass[i, j] = 255
# 创建一个与输入图像大小相同的空白图像 # 应用滤波器到频域表示
Ideal_LowPass = np.zeros((rows, cols), dtype=np.uint8) mask = Ideal_LowPass[:, :, np.newaxis]
# 创建理想低通滤波器 fshift = dft_shift * mask
for i in range(rows): # 逆傅里叶变换以获得平滑后的图像
for j in range(cols): f_ishift = np.fft.ifftshift(fshift)
x = i - crow img_back = cv2.idft(f_ishift)
y = j - ccol img_back = cv2.magnitude(img_back[:, :, 0], img_back[:, :, 1])
D = np.sqrt(x**2 + y**2) # 归一化图像到0-255
if D <= D0: cv2.normalize(img_back, img_back, 0, 255, cv2.NORM_MINMAX)
Ideal_LowPass[i, j] = 255 img_back = np.uint8(img_back)
# 应用滤波器到频域表示
mask = Ideal_LowPass[:, :, np.newaxis] return img_back
fshift = dft_shift * mask
# 逆傅里叶变换以获得平滑后的图像 # 布特沃斯低通滤波器
f_ishift = np.fft.ifftshift(fshift) def ButterWorth_LowPassFilter(rows, cols, crow, ccol, D0=20, n=2):
img_back = cv2.idft(f_ishift) # 创建一个与输入图像大小相同的空白图像
img_back = cv2.magnitude(img_back[:, :, 0], img_back[:, :, 1]) ButterWorth_LowPass = np.zeros((rows, cols), dtype=np.uint8)
# 归一化图像到0-255 # 创建巴特沃斯低通滤波器
cv2.normalize(img_back, img_back, 0, 255, cv2.NORM_MINMAX) for i in range(rows):
img_back = np.uint8(img_back) for j in range(cols):
x = i - crow
return img_back y = j - ccol
D = np.sqrt(x ** 2 + y ** 2)
# 布特沃斯低通滤波器 ButterWorth_LowPass[i, j] = 255 / (1 + (D / D0) ** (2 * n))
def ButterWorth_LowPassFilter(rows, cols, crow, ccol, D0=20, n=2): # 应用滤波器到频域表示
# 创建一个与输入图像大小相同的空白图像 mask = ButterWorth_LowPass[:, :, np.newaxis]
ButterWorth_LowPass = np.zeros((rows, cols), dtype=np.uint8) fshift = dft_shift * mask
# 创建巴特沃斯低通滤波器 # 逆傅里叶变换以获得平滑后的图像
for i in range(rows): f_ishift = np.fft.ifftshift(fshift)
for j in range(cols): img_back = cv2.idft(f_ishift)
x = i - crow img_back = cv2.magnitude(img_back[:, :, 0], img_back[:, :, 1])
y = j - ccol # 归一化图像到0-255
D = np.sqrt(x ** 2 + y ** 2) cv2.normalize(img_back, img_back, 0, 255, cv2.NORM_MINMAX)
ButterWorth_LowPass[i, j] = 255 / (1 + (D / D0) ** (2 * n)) img_back = np.uint8(img_back)
# 应用滤波器到频域表示
mask = ButterWorth_LowPass[:, :, np.newaxis] return img_back
fshift = dft_shift * mask
# 逆傅里叶变换以获得平滑后的图像 # 高斯低通滤波器
f_ishift = np.fft.ifftshift(fshift) def Gauss_LowPassFilter(rows, cols, crow, ccol, D0=20):
img_back = cv2.idft(f_ishift) # 创建一个与输入图像大小相同的空白图像
img_back = cv2.magnitude(img_back[:, :, 0], img_back[:, :, 1]) Gauss_LowPass = np.zeros((rows, cols), dtype=np.uint8)
# 归一化图像到0-255 # 创建高斯低通滤波器
cv2.normalize(img_back, img_back, 0, 255, cv2.NORM_MINMAX) for i in range(rows):
img_back = np.uint8(img_back) for j in range(cols):
x = i - crow
return img_back y = j - ccol
D = np.sqrt(x ** 2 + y ** 2)
# 高斯低通滤波器 Gauss_LowPass[i, j] = 255 * np.exp(-0.5 * (D ** 2) / (D0 ** 2))
def Gauss_LowPassFilter(rows, cols, crow, ccol, D0=20): # 应用滤波器到频域表示
# 创建一个与输入图像大小相同的空白图像 mask = Gauss_LowPass[:, :, np.newaxis]
Gauss_LowPass = np.zeros((rows, cols), dtype=np.uint8) fshift = dft_shift * mask
# 创建高斯低通滤波器 # 逆傅里叶变换以获得平滑后的图像
for i in range(rows): f_ishift = np.fft.ifftshift(fshift)
for j in range(cols): img_back = cv2.idft(f_ishift)
x = i - crow img_back = cv2.magnitude(img_back[:, :, 0], img_back[:, :, 1])
y = j - ccol # 归一化图像到0-255
D = np.sqrt(x ** 2 + y ** 2) cv2.normalize(img_back, img_back, 0, 255, cv2.NORM_MINMAX)
Gauss_LowPass[i, j] = 255 * np.exp(-0.5 * (D ** 2) / (D0 ** 2)) img_back = np.uint8(img_back)
# 应用滤波器到频域表示
mask = Gauss_LowPass[:, :, np.newaxis] return img_back
fshift = dft_shift * mask
# 逆傅里叶变换以获得平滑后的图像 # 读取灰度图像
f_ishift = np.fft.ifftshift(fshift) im = cv2.cvtColor(src, cv2.COLOR_BGR2GRAY)
img_back = cv2.idft(f_ishift)
img_back = cv2.magnitude(img_back[:, :, 0], img_back[:, :, 1]) # 获取图像的频域表示
# 归一化图像到0-255 dft = cv2.dft(np.float32(im), flags=cv2.DFT_COMPLEX_OUTPUT)
cv2.normalize(img_back, img_back, 0, 255, cv2.NORM_MINMAX) dft_shift = np.fft.fftshift(dft)
img_back = np.uint8(img_back)
# 获取图像的尺寸
return img_back rows, cols = im.shape
crow, ccol = rows // 2, cols // 2
# 读取灰度图像
im = cv2.cvtColor(src, cv2.COLOR_BGR2GRAY) # 理想低通滤波器
Ideal_LowPass = Ideal_LowPassFilter(rows, cols, crow, ccol)
# 获取图像的频域表示 # 巴特沃斯低通滤波器
dft = cv2.dft(np.float32(im), flags=cv2.DFT_COMPLEX_OUTPUT) ButterWorth_LowPass = ButterWorth_LowPassFilter(rows, cols, crow, ccol)
dft_shift = np.fft.fftshift(dft) # 高斯低通滤波器
Gauss_LowPass = Gauss_LowPassFilter(rows, cols, crow, ccol)
# 获取图像的尺寸
rows, cols = im.shape combined = np.hstack((Ideal_LowPass, ButterWorth_LowPass, Gauss_LowPass))
crow, ccol = rows // 2, cols // 2 # 更新 edge 变量
edge = Image.fromarray(combined)
# 理想低通滤波器
Ideal_LowPass = Ideal_LowPassFilter(rows, cols, crow, ccol) # 创建Toplevel窗口
# 巴特沃斯低通滤波器 try:
ButterWorth_LowPass = ButterWorth_LowPassFilter(rows, cols, crow, ccol) FreqsmoWin.destroy()
# 高斯低通滤波器 except Exception as e:
Gauss_LowPass = Gauss_LowPassFilter(rows, cols, crow, ccol) print("NVM")
finally:
combined = np.hstack((Ideal_LowPass, ButterWorth_LowPass, Gauss_LowPass)) FreqsmoWin = Toplevel()
# 更新 edge 变量 FreqsmoWin.attributes('-topmost', True)
edge = Image.fromarray(combined) FreqsmoWin.geometry("720x300")
FreqsmoWin.resizable(True, True) # 可缩放
# 创建Toplevel窗口 FreqsmoWin.title("频域平滑结果")
try:
FreqsmoWin.destroy() # 显示图像
except Exception as e: LabelPic = tk.Label(FreqsmoWin, text="IMG", width=720, height=240)
print("NVM") image = ImageTk.PhotoImage(Image.fromarray(combined))
finally: LabelPic.image = image
FreqsmoWin = Toplevel() LabelPic['image'] = image
FreqsmoWin.attributes('-topmost', True)
FreqsmoWin.geometry("720x300") LabelPic.bind('<Configure>', lambda event: changeSize(event, combined, LabelPic))
FreqsmoWin.resizable(True, True) # 可缩放 LabelPic.pack(fill=tk.BOTH, expand=tk.YES)
FreqsmoWin.title("频域平滑结果")
# 添加保存按钮
# 显示图像 btn_save = tk.Button(FreqsmoWin, text="保存", bg='#add8e6', fg='black', font=('Helvetica', 14), width=20,
LabelPic = tk.Label(FreqsmoWin, text="IMG", width=720, height=240) command=savefile)
image = ImageTk.PhotoImage(Image.fromarray(combined)) btn_save.pack(pady=10)
LabelPic.image = image
LabelPic['image'] = image return
LabelPic.bind('<Configure>', lambda event: changeSize(event, combined, LabelPic)) #空域平滑
LabelPic.pack(fill=tk.BOTH, expand=tk.YES) def air_smo(root):
global src, AirsmoWin, edge
# 添加保存按钮
btn_save = tk.Button(FreqsmoWin, text="保存", bg='#add8e6', fg='black', font=('Helvetica', 14), width=20, # 判断是否已经选取图片
command=savefile) if src is None:
btn_save.pack(pady=10) messagebox.showerror("错误", "没有选择图片！")
return
return
# 均值平滑滤波
#空域平滑 def mean_filter(image, height, width):
def air_smo(root): # 创建空白图像以存储滤波结果
global src, AirsmoWin, edge filtered_image = np.zeros((height - 2, width - 2), dtype=np.uint8)
# 判断是否已经选取图片 # 执行3x3均值滤波
if src is None: for i in range(1, height - 1):
messagebox.showerror("错误", "没有选择图片！") for j in range(1, width - 1):
return tmp = (int(image[i - 1, j - 1]) + int(image[i - 1, j]) + int(image[i - 1, j + 1]) +
int(image[i, j - 1]) + int(image[i, j]) + int(image[i, j + 1]) +
# 均值平滑滤波 int(image[i + 1, j - 1]) + int(image[i + 1, j]) + int(image[i + 1, j + 1])) // 9
def mean_filter(image, height, width): filtered_image[i - 1, j - 1] = tmp
# 创建空白图像以存储滤波结果
filtered_image = np.zeros((height - 2, width - 2), dtype=np.uint8) return filtered_image
# 执行3x3均值滤波 # 中值平滑滤波
for i in range(1, height - 1): def median_filter(image, height, width):
for j in range(1, width - 1): # 创建空白图像以存储滤波结果
tmp = (int(image[i - 1, j - 1]) + int(image[i - 1, j]) + int(image[i - 1, j + 1]) + filtered_image = np.zeros((height - 2, width - 2), dtype=np.uint8)
int(image[i, j - 1]) + int(image[i, j]) + int(image[i, j + 1]) +
int(image[i + 1, j - 1]) + int(image[i + 1, j]) + int(image[i + 1, j + 1])) // 9 # 执行3x3中值滤波
filtered_image[i - 1, j - 1] = tmp for i in range(1, height - 1):
for j in range(1, width - 1):
return filtered_image # 取3x3邻域
region = [
# 中值平滑滤波 image[i - 1, j - 1], image[i - 1, j], image[i - 1, j + 1],
def median_filter(image, height, width): image[i, j - 1], image[i, j], image[i, j + 1],
# 创建空白图像以存储滤波结果 image[i + 1, j - 1], image[i + 1, j], image[i + 1, j + 1]
filtered_image = np.zeros((height - 2, width - 2), dtype=np.uint8) ]
# 计算中值
# 执行3x3中值滤波 filtered_image[i - 1, j - 1] = np.median(region)
for i in range(1, height - 1):
for j in range(1, width - 1): return filtered_image
# 取3x3邻域
region = [ # 5x5 中值平滑滤波
image[i - 1, j - 1], image[i - 1, j], image[i - 1, j + 1], def med_filter_5x5(image, height, width):
image[i, j - 1], image[i, j], image[i, j + 1], # 创建空白图像以存储滤波结果
image[i + 1, j - 1], image[i + 1, j], image[i + 1, j + 1] filtered_image = np.zeros((height - 4, width - 4), dtype=np.uint8)
]
# 计算中值 # 执行5x5中值滤波
filtered_image[i - 1, j - 1] = np.median(region) for i in range(2, height - 2):
for j in range(2, width - 2):
return filtered_image # 取5x5邻域的所有值
neighbors = [
# 5x5 中值平滑滤波 image[i - 2, j - 2], image[i - 2, j - 1], image[i - 2, j], image[i - 2, j + 1], image[i - 2, j + 2],
def med_filter_5x5(image, height, width): image[i - 1, j - 2], image[i - 1, j - 1], image[i - 1, j], image[i - 1, j + 1], image[i - 1, j + 2],
# 创建空白图像以存储滤波结果 image[i, j - 2], image[i, j - 1], image[i, j], image[i, j + 1], image[i, j + 2],
filtered_image = np.zeros((height - 4, width - 4), dtype=np.uint8) image[i + 1, j - 2], image[i + 1, j - 1], image[i + 1, j], image[i + 1, j + 1], image[i + 1, j + 2],
image[i + 2, j - 2], image[i + 2, j - 1], image[i + 2, j], image[i + 2, j + 1], image[i + 2, j + 2]
# 执行5x5中值滤波 ]
for i in range(2, height - 2): # 计算中值
for j in range(2, width - 2): filtered_image[i - 2, j - 2] = np.median(neighbors)
# 取5x5邻域的所有值
neighbors = [ return filtered_image
image[i - 2, j - 2], image[i - 2, j - 1], image[i - 2, j], image[i - 2, j + 1], image[i - 2, j + 2],
image[i - 1, j - 2], image[i - 1, j - 1], image[i - 1, j], image[i - 1, j + 1], image[i - 1, j + 2], # 读取灰度图像
image[i, j - 2], image[i, j - 1], image[i, j], image[i, j + 1], image[i, j + 2], im = cv2.cvtColor(src, cv2.COLOR_BGR2GRAY)
image[i + 1, j - 2], image[i + 1, j - 1], image[i + 1, j], image[i + 1, j + 1], image[i + 1, j + 2],
image[i + 2, j - 2], image[i + 2, j - 1], image[i + 2, j], image[i + 2, j + 1], image[i + 2, j + 2] # 获取图像尺寸
] height, width = im.shape
# 计算中值
filtered_image[i - 2, j - 2] = np.median(neighbors) # 邻域平均
mean = mean_filter(im, height, width)
return filtered_image # 中值滤波3x3
median = median_filter(im, height, width)
# 读取灰度图像 # 中值滤波5x5
im = cv2.cvtColor(src, cv2.COLOR_BGR2GRAY) med = med_filter_5x5(im, height, width)
# 获取图像尺寸 min_height = min(mean.shape[0], median.shape[0], med.shape[0])
height, width = im.shape min_width = min(mean.shape[1], median.shape[1], med.shape[1])
# 邻域平均 mean_cropped = mean[:min_height, :min_width]
mean = mean_filter(im, height, width) median_cropped = median[:min_height, :min_width]
# 中值滤波3x3 med_cropped = med[:min_height, :min_width]
median = median_filter(im, height, width)
# 中值滤波5x5 combined = np.hstack((mean_cropped, median_cropped, med_cropped))
med = med_filter_5x5(im, height, width) # 更新 edge 变量
edge = Image.fromarray(combined)
min_height = min(mean.shape[0], median.shape[0], med.shape[0])
min_width = min(mean.shape[1], median.shape[1], med.shape[1]) # 创建Toplevel窗口
try:
mean_cropped = mean[:min_height, :min_width] AirsmoWin.destroy()
median_cropped = median[:min_height, :min_width] except Exception as e:
med_cropped = med[:min_height, :min_width] print("NVM")
finally:
combined = np.hstack((mean_cropped, median_cropped, med_cropped)) AirsmoWin = Toplevel()
# 更新 edge 变量 AirsmoWin.attributes('-topmost', True)
edge = Image.fromarray(combined) AirsmoWin.geometry("720x300")
AirsmoWin.resizable(True, True) # 可缩放
# 创建Toplevel窗口 AirsmoWin.title("空域平滑结果")
try:
AirsmoWin.destroy() # 显示图像
except Exception as e: LabelPic = tk.Label(AirsmoWin, text="IMG", width=720, height=240)
print("NVM") image = ImageTk.PhotoImage(Image.fromarray(combined))
finally: LabelPic.image = image
AirsmoWin = Toplevel() LabelPic['image'] = image
AirsmoWin.attributes('-topmost', True)
AirsmoWin.geometry("720x300") LabelPic.bind('<Configure>', lambda event: changeSize(event, combined, LabelPic))
AirsmoWin.resizable(True, True) # 可缩放 LabelPic.pack(fill=tk.BOTH, expand=tk.YES)
AirsmoWin.title("空域平滑结果")
# 添加保存按钮
# 显示图像 btn_save = tk.Button(AirsmoWin, text="保存", bg='#add8e6', fg='black', font=('Helvetica', 14), width=20,
LabelPic = tk.Label(AirsmoWin, text="IMG", width=720, height=240) command=savefile)
image = ImageTk.PhotoImage(Image.fromarray(combined)) btn_save.pack(pady=10)
LabelPic.image = image
LabelPic['image'] = image
LabelPic.bind('<Configure>', lambda event: changeSize(event, combined, LabelPic))
LabelPic.pack(fill=tk.BOTH, expand=tk.YES)
# 添加保存按钮
btn_save = tk.Button(AirsmoWin, text="保存", bg='#add8e6', fg='black', font=('Helvetica', 14), width=20,
command=savefile)
btn_save.pack(pady=10)
return return
Write Preview
Loading…
Cancel
Save