Image_Process/basic/image_smooth.py

import tkinter as tk
from tkinter import filedialog, messagebox
from tkinter import Toplevel
from PIL import Image, ImageTk
import numpy as np
import cv2
import os

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
img_label = None  # 全局变量，用于存储显示选择的图片的标签
edge = None

FreqsmoWin = None
AirsmoWin  = None

def select_image(root):
    global img_path, src, img_label, edge

    img_path = filedialog.askopenfilename(filetypes=[("Image files", "*.jpg;*.png;*.jpeg;*.bmp")])
    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
        img_path_fixed = os.path.normpath(img_path)

        # 图像输入
        src_temp = cv2.imdecode(np.fromfile(img_path_fixed, dtype=np.uint8), cv2.IMREAD_UNCHANGED)
        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")
            return
        src = cv2.cvtColor(src_temp, cv2.COLOR_BGR2RGB)

        # 检查 img_label 是否存在且有效
        if img_label is None or not img_label.winfo_exists():
            img_label = tk.Label(root)
            img_label.pack(side=tk.TOP, pady=10)

        img = Image.open(img_path)
        img.thumbnail((160, 160))
        img_tk = ImageTk.PhotoImage(img)
        img_label.configure(image=img_tk)
        img_label.image = img_tk

        # 定义 edge 变量为 PIL.Image 对象，以便稍后保存
        edge = Image.fromarray(src)
    else:
        messagebox.showerror("错误", "没有选择图片路径")

def show_selected_image(root):
    global img_label
    img_label = tk.Label(root)
    img_label.pack(side=tk.TOP, pady=10)
    img = Image.open(img_path)
    img.thumbnail((160, 160))
    img_tk = ImageTk.PhotoImage(img)
    img_label.configure(image=img_tk)
    img_label.image = img_tk

def changeSize(event, img, LabelPic):
    img_aspect = img.shape[1] / img.shape[0]
    new_aspect = event.width / event.height

    if new_aspect > img_aspect:
        new_width = int(event.height * img_aspect)
        new_height = event.height
    else:
        new_width = event.width
        new_height = int(event.width / img_aspect)

    resized_image = cv2.resize(img, (new_width, new_height))
    image1 = ImageTk.PhotoImage(Image.fromarray(resized_image))
    LabelPic.image = image1
    LabelPic['image'] = image1

def savefile():
    global edge

    filename = filedialog.asksaveasfilename(defaultextension=".jpg", filetypes=[("JPEG files", "*.jpg"), ("PNG files", "*.png"), ("BMP files", "*.bmp")])
    if not filename:
        return
    # 确保 edge 变量已定义
    if edge is not None:
        try:
            edge.save(filename)
            messagebox.showinfo("保存成功", "图片保存成功！")
        except Exception as e:
            messagebox.showerror("保存失败", f"无法保存图片: {e}")
    else:
        messagebox.showerror("保存失败", "没有图像可保存")

#频域平滑
def freq_smo(root):
    global src, FreqsmoWin, edge

    # 判断是否已经选取图片
    if src is None:
        messagebox.showerror("错误", "没有选择图片！")
        return

    # 理想低通滤波器
    def Ideal_LowPassFilter(rows, cols, crow, ccol, D0=20):
        # 创建一个与输入图像大小相同的空白图像
        Ideal_LowPass = np.zeros((rows, cols), dtype=np.uint8)
        # 创建理想低通滤波器
        for i in range(rows):
            for j in range(cols):
                x = i - crow
                y = j - ccol
                D = np.sqrt(x**2 + y**2)
                if D <= D0:
                    Ideal_LowPass[i, j] = 255
        # 应用滤波器到频域表示
        mask = Ideal_LowPass[:, :, np.newaxis]
        fshift = dft_shift * mask
        # 逆傅里叶变换以获得平滑后的图像
        f_ishift = np.fft.ifftshift(fshift)
        img_back = cv2.idft(f_ishift)
        img_back = cv2.magnitude(img_back[:, :, 0], img_back[:, :, 1])
        # 归一化图像到0-255
        cv2.normalize(img_back, img_back, 0, 255, cv2.NORM_MINMAX)
        img_back = np.uint8(img_back)

        return img_back

    # 布特沃斯低通滤波器
    def ButterWorth_LowPassFilter(rows, cols, crow, ccol, D0=20, n=2):
        # 创建一个与输入图像大小相同的空白图像
        ButterWorth_LowPass = np.zeros((rows, cols), dtype=np.uint8)
        # 创建巴特沃斯低通滤波器
        for i in range(rows):
            for j in range(cols):
                x = i - crow
                y = j - ccol
                D = np.sqrt(x ** 2 + y ** 2)
                ButterWorth_LowPass[i, j] = 255 / (1 + (D / D0) ** (2 * n))
        # 应用滤波器到频域表示
        mask = ButterWorth_LowPass[:, :, np.newaxis]
        fshift = dft_shift * mask
        # 逆傅里叶变换以获得平滑后的图像
        f_ishift = np.fft.ifftshift(fshift)
        img_back = cv2.idft(f_ishift)
        img_back = cv2.magnitude(img_back[:, :, 0], img_back[:, :, 1])
        # 归一化图像到0-255
        cv2.normalize(img_back, img_back, 0, 255, cv2.NORM_MINMAX)
        img_back = np.uint8(img_back)

        return img_back

    # 高斯低通滤波器
    def Gauss_LowPassFilter(rows, cols, crow, ccol, D0=20):
        # 创建一个与输入图像大小相同的空白图像
        Gauss_LowPass = np.zeros((rows, cols), dtype=np.uint8)
        # 创建高斯低通滤波器
        for i in range(rows):
            for j in range(cols):
                x = i - crow
                y = j - ccol
                D = np.sqrt(x ** 2 + y ** 2)
                Gauss_LowPass[i, j] = 255 * np.exp(-0.5 * (D ** 2) / (D0 ** 2))
        # 应用滤波器到频域表示
        mask = Gauss_LowPass[:, :, np.newaxis]
        fshift = dft_shift * mask
        # 逆傅里叶变换以获得平滑后的图像
        f_ishift = np.fft.ifftshift(fshift)
        img_back = cv2.idft(f_ishift)
        img_back = cv2.magnitude(img_back[:, :, 0], img_back[:, :, 1])
        # 归一化图像到0-255
        cv2.normalize(img_back, img_back, 0, 255, cv2.NORM_MINMAX)
        img_back = np.uint8(img_back)

        return img_back

    # 读取灰度图像
    im = cv2.cvtColor(src, cv2.COLOR_BGR2GRAY)

    # 获取图像的频域表示
    dft = cv2.dft(np.float32(im), flags=cv2.DFT_COMPLEX_OUTPUT)
    dft_shift = np.fft.fftshift(dft)

    # 获取图像的尺寸
    rows, cols = im.shape
    crow, ccol = rows // 2, cols // 2

    # 理想低通滤波器
    Ideal_LowPass = Ideal_LowPassFilter(rows, cols, crow, ccol)
    # 巴特沃斯低通滤波器
    ButterWorth_LowPass = ButterWorth_LowPassFilter(rows, cols, crow, ccol)
    # 高斯低通滤波器
    Gauss_LowPass = Gauss_LowPassFilter(rows, cols, crow, ccol)

    combined = np.hstack((Ideal_LowPass, ButterWorth_LowPass, Gauss_LowPass))
    # 更新 edge 变量
    edge = Image.fromarray(combined)

    # 创建Toplevel窗口
    try:
        FreqsmoWin.destroy()
    except Exception as e:
        print("NVM")
    finally:
        FreqsmoWin = Toplevel()
        FreqsmoWin.attributes('-topmost', True)
    FreqsmoWin.geometry("720x300")
    FreqsmoWin.resizable(True, True)  # 可缩放
    FreqsmoWin.title("频域平滑结果")

    # 显示图像
    LabelPic = tk.Label(FreqsmoWin, text="IMG", width=720, height=240)
    image = ImageTk.PhotoImage(Image.fromarray(combined))
    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(FreqsmoWin, text="保存", bg='#add8e6', fg='black', font=('Helvetica', 14), width=20,
                         command=savefile)
    btn_save.pack(pady=10)

    return

#空域平滑
def air_smo(root):
    global src, AirsmoWin, edge

    # 判断是否已经选取图片
    if src is None:
        messagebox.showerror("错误", "没有选择图片！")
        return

    # 均值平滑滤波
    def mean_filter(image, height, width):
        # 创建空白图像以存储滤波结果
        filtered_image = np.zeros((height - 2, width - 2), dtype=np.uint8)

        # 执行3x3均值滤波
        for i in range(1, height - 1):
            for j in range(1, width - 1):
                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]) +
                       int(image[i + 1, j - 1]) + int(image[i + 1, j]) + int(image[i + 1, j + 1])) // 9
                filtered_image[i - 1, j - 1] = tmp

        return filtered_image

    # 中值平滑滤波
    def median_filter(image, height, width):
        # 创建空白图像以存储滤波结果
        filtered_image = np.zeros((height - 2, width - 2), dtype=np.uint8)

        # 执行3x3中值滤波
        for i in range(1, height - 1):
            for j in range(1, width - 1):
                # 取3x3邻域
                region = [
                    image[i - 1, j - 1], image[i - 1, j], image[i - 1, j + 1],
                    image[i, j - 1], image[i, j], image[i, j + 1],
                    image[i + 1, j - 1], image[i + 1, j], image[i + 1, j + 1]
                ]
                # 计算中值
                filtered_image[i - 1, j - 1] = np.median(region)

        return filtered_image

    # 5x5 中值平滑滤波
    def med_filter_5x5(image, height, width):
        # 创建空白图像以存储滤波结果
        filtered_image = np.zeros((height - 4, width - 4), dtype=np.uint8)

        # 执行5x5中值滤波
        for i in range(2, height - 2):
            for j in range(2, width - 2):
                # 取5x5邻域的所有值
                neighbors = [
                    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],
                    image[i, j - 2], image[i, j - 1], image[i, j], image[i, j + 1], image[i, 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],
                    image[i + 2, j - 2], image[i + 2, j - 1], image[i + 2, j], image[i + 2, j + 1], image[i + 2, j + 2]
                ]
                # 计算中值
                filtered_image[i - 2, j - 2] = np.median(neighbors)

        return filtered_image

    # 读取灰度图像
    im = cv2.cvtColor(src, cv2.COLOR_BGR2GRAY)

    # 获取图像尺寸
    height, width = im.shape

    # 邻域平均
    mean = mean_filter(im, height, width)
    # 中值滤波3x3
    median = median_filter(im, height, width)
    # 中值滤波5x5
    med = med_filter_5x5(im, height, width)

    min_height = min(mean.shape[0], median.shape[0], med.shape[0])
    min_width = min(mean.shape[1], median.shape[1], med.shape[1])

    mean_cropped = mean[:min_height, :min_width]
    median_cropped = median[:min_height, :min_width]
    med_cropped = med[:min_height, :min_width]

    combined = np.hstack((mean_cropped, median_cropped, med_cropped))
    # 更新 edge 变量
    edge = Image.fromarray(combined)

    # 创建Toplevel窗口
    try:
        AirsmoWin.destroy()
    except Exception as e:
        print("NVM")
    finally:
        AirsmoWin = Toplevel()
        AirsmoWin.attributes('-topmost', True)
    AirsmoWin.geometry("720x300")
    AirsmoWin.resizable(True, True)  # 可缩放
    AirsmoWin.title("空域平滑结果")

    # 显示图像
    LabelPic = tk.Label(AirsmoWin, text="IMG", width=720, height=240)
    image = ImageTk.PhotoImage(Image.fromarray(combined))
    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