Python使用Pillow与OpenCV进行图像处理实战

pip install pillow

from PIL import Image, ImageDraw, ImageFont, ImageFilter, ImageEnhance

# 图片尺寸 = 宽 × 高 (单位：像素)
# 1920 × 1080 = 2,073,600 像素 (约200万像素，即 2MP)

# DPI (Dots Per Inch) = 每英寸的像素数
# 72 DPI: 屏幕显示标准
# 150 DPI: 普通打印
# 300 DPI: 高质量打印

from PIL import Image

# 打开图片
img = Image.open('photo.jpg')

# 查看图片信息
print(f"格式: {img.format}")      # JPEG, PNG 等
print(f"模式: {img.mode}")        # RGB, RGBA, L 等
print(f"尺寸: {img.size}")        # (宽度, 高度)
print(f"宽度: {img.width} 像素")
print(f"高度: {img.height} 像素")

# 查看 EXIF 信息（相机参数等）
exif = img._getexif()
if exif:
    for tag_id, value in exif.items():
        print(f"  {tag_id}: {value}")

# 创建空白图片
img = Image.new('RGB', (800, 400), color='#2c3e50')

# 创建透明图片
img = Image.new('RGBA', (400, 300), (0, 0, 0, 0))

# 从像素数据创建
pixels = [(255, 0, 0), (0, 255, 0), (0, 0, 255)]
img = Image.new('RGB', (3, 1))
img.putdata(pixels)

img = Image.open('source.png')

# 保存为不同格式
img.save('output.jpg', 'JPEG', quality=95)   # JPEG, 质量95%
img.save('output.bmp', 'BMP')                # BMP 无损
img.save('output.webp', 'WEBP')              # WebP 现代格式
img.save('output.gif', 'GIF')                # GIF
img.save('output.tiff', 'TIFF')              # TIFF 专业格式

格式测试输出:
  PNG: .png → 1.9 KB     ← 无损压缩，体积中等
  JPEG: .jpg → 6.5 KB    ← 有损压缩，体积较大
  BMP: .bmp → 351.6 KB   ← 无压缩，体积巨大
  WEBP: .webp → 0.9 KB   ← 新一代格式，体积最小

img = Image.open('photo.jpg')

# 裁剪区域: (左, 上, 右, 下)
cropped = img.crop((100, 50, 500, 350))
cropped.save('cropped.jpg')

# 指定尺寸缩放（可能变形）
resized = img.resize((400, 300))

# 高质量缩放（使用 LANCZOS 滤波器）
resized = img.resize((400, 300), Image.LANCZOS)

# 等比缩放（保持宽高比）
img.thumbnail((800, 800))  # 最大边不超过800

# 放大 2 倍
enlarged = img.resize((img.width * 2, img.height * 2), Image.LANCZOS)

img = Image.open('photo.jpg')

# 旋转
rotated_45 = img.rotate(45)                    # 逆时针旋转45度
rotated_90 = img.rotate(90, expand=True)       # 旋转90度，扩展画布
rotated_free = img.rotate(30, expand=True, fillcolor='#2c3e50')  # 自定义背景色

# 翻转
flipped_h = img.transpose(Image.FLIP_LEFT_RIGHT)  # 水平翻转
flipped_v = img.transpose(Image.FLIP_TOP_BOTTOM)  # 垂直翻转

from PIL import Image, ImageDraw, ImageFont

img = Image.new('RGB', (800, 400), '#ecf0f1')
draw = ImageDraw.Draw(img)

# 绘制矩形
draw.rectangle([50, 50, 200, 200], fill='#e74c3c', outline='white', width=3)

# 绘制椭圆
draw.ellipse([250, 50, 400, 200], fill='#3498db', outline='white', width=3)

# 绘制三角形
draw.polygon([(500, 200), (450, 50), (550, 50)], fill='#2ecc71', outline='white')

# 绘制圆角矩形
draw.rounded_rectangle([600, 50, 750, 200], radius=20, fill='#f39c12')

# 绘制线条
draw.line([(50, 250), (750, 250)], fill='#2c3e50', width=3)

# 绘制文字
font = ImageFont.truetype("msyh.ttc", 36)  # 微软雅黑36号
draw.text((200, 300), 'Hello Pillow!', fill='#2c3e50', font=font)

from PIL import ImageFilter

img = Image.open('photo.jpg')

# 模糊
img_blur = img.filter(ImageFilter.GaussianBlur(radius=5))
img_blur = img.filter(ImageFilter.BLUR)           # 默认模糊

# 锐化
img_sharpen = img.filter(ImageFilter.SHARPEN)
img_sharpen = img.filter(ImageFilter.UnsharpMask(radius=2, percent=150))

# 边缘检测
img_edges = img.filter(ImageFilter.FIND_EDGES)

# 浮雕
img_emboss = img.filter(ImageFilter.EMBOSS)

# 轮廓
img_contour = img.filter(ImageFilter.CONTOUR)

# 细节增强
img_detail = img.filter(ImageFilter.DETAIL)

# 平滑
img_smooth = img.filter(ImageFilter.SMOOTH)
img_smooth_more = img.filter(ImageFilter.SMOOTH_MORE)

# 多重滤镜叠加
result = img.filter(ImageFilter.GaussianBlur(2)).filter(ImageFilter.SHARPEN)

from PIL import ImageEnhance

img = Image.open('photo.jpg')

# 亮度调整 (1.0=原始, >1增强, <1减弱)
enhancer = ImageEnhance.Brightness(img)
bright = enhancer.enhance(1.5)  # 亮度提高50%

# 对比度调整
enhancer = ImageEnhance.Contrast(img)
contrast = enhancer.enhance(1.8)  # 对比度提高80%

# 色彩饱和度调整
enhancer = ImageEnhance.Color(img)
color = enhancer.enhance(2.0)  # 饱和度翻倍

# 锐度调整
enhancer = ImageEnhance.Sharpness(img)
sharp = enhancer.enhance(2.0)  # 锐度翻倍

# 综合调整
from PIL import ImageOps
img_auto = ImageOps.autocontrast(img)    # 自动对比度
img_equalize = ImageOps.equalize(img)    # 直方图均衡化
img_invert = ImageOps.invert(img.convert('RGB'))  # 反色
img_gray = ImageOps.grayscale(img)       # 转灰度

from PIL import Image, ImageDraw, ImageFont

def add_text_watermark(image_path, text, output_path, opacity=128):
    img = Image.open(image_path).convert('RGBA')
    watermark = Image.new('RGBA', img.size, (0, 0, 0, 0))
    draw = ImageDraw.Draw(watermark)

    font = ImageFont.truetype("msyh.ttc", 36)
    # 在右下角添加水印
    bbox = draw.textbbox((0, 0), text, font=font)
    x = img.width - bbox[2] - 20
    y = img.height - bbox[3] - 20
    draw.text((x, y), text, fill=(128, 128, 128, opacity), font=font)

    result = Image.alpha_composite(img, watermark)
    result.convert('RGB').save(output_path)

def add_image_watermark(image_path, watermark_path, output_path, position='bottom-right'):
    img = Image.open(image_path).convert('RGBA')
    watermark = Image.open(watermark_path).convert('RGBA')

    # 缩放水印
    wm_size = (img.width // 5, img.height // 5)
    watermark = watermark.resize(wm_size, Image.LANCZOS)

    # 计算位置
    if position == 'bottom-right':
        x = img.width - wm_size[0] - 20
        y = img.height - wm_size[1] - 20
    elif position == 'center':
        x = (img.width - wm_size[0]) // 2
        y = (img.height - wm_size[1]) // 2

    # 调整透明度
    watermark.putalpha(watermark.split()[3])

    # 合成
    img.paste(watermark, (x, y), watermark)
    img.convert('RGB').save(output_path)

def combine_images(images, cols=2, padding=10, bg_color='#2c3e50'):
    rows_count = (len(images) + cols - 1) // cols
    max_w = max(img.width for img in images)
    max_h = max(img.height for img in images)

    combined_w = cols * max_w + (cols + 1) * padding
    combined_h = rows_count * max_h + (rows_count + 1) * padding

    combined = Image.new('RGB', (combined_w, combined_h), bg_color)

    for i, img in enumerate(images):
        row = i // cols
        col = i % cols
        x = padding + col * (max_w + padding)
        y = padding + row * (max_h + padding)
        combined.paste(img, (x, y))

    return combined

"""
批量处理图片：调整大小 + 加水印 + 转换格式
"""
from PIL import Image, ImageDraw, ImageFont, ImageEnhance, ImageFilter
import os

def batch_process(input_dir, output_dir, max_size=(1920, 1080),
                  watermark_text=None, output_format='JPEG'):
    """批量处理图片"""
    os.makedirs(output_dir, exist_ok=True)

    supported = {'.jpg', '.jpeg', '.png', '.bmp', '.webp', '.tiff'}

    for filename in os.listdir(input_dir):
        ext = os.path.splitext(filename)[1].lower()
        if ext not in supported:
            continue

        filepath = os.path.join(input_dir, filename)
        try:
            img = Image.open(filepath)
        except Exception:
            print(f"跳过无法打开的文件: {filename}")
            continue

        # 1. 转为 RGB（去除透明通道）
        if img.mode in ('RGBA', 'P'):
            img = img.convert('RGB')

        # 2. 缩放到最大尺寸
        img.thumbnail(max_size, Image.LANCZOS)

        # 3. 轻微锐化
        img = img.filter(ImageFilter.SHARPEN)

        # 4. 自动对比度
        from PIL import ImageOps
        img = ImageOps.autocontrast(img)

        # 5. 添加水印
        if watermark_text:
            draw = ImageDraw.Draw(img)
            try:
                font = ImageFont.truetype("msyh.ttc", max(20, img.width // 30))
            except:
                font = ImageFont.load_default()
            text = watermark_text
            bbox = draw.textbbox((0, 0), text, font=font)
            x = img.width - (bbox[2] - bbox[0]) - 20
            y = img.height - (bbox[3] - bbox[1]) - 20
            draw.text((x, y), text, fill=(255, 255, 255), font=font)

        # 6. 保存
        name = os.path.splitext(filename)[0]
        output_path = os.path.join(output_dir, f"{name}.{output_format.lower()}")
        img.save(output_path, output_format, quality=90)
        print(f"已处理: {filename} → {output_path}")

# 使用
batch_process(
    input_dir='./photos',
    output_dir='./photos_processed',
    max_size=(1920, 1080),
    watermark_text='© 2025 MyBlog'
)

pip install opencv-python
pip install opencv-contrib-python  # 包含额外模块

import cv2
import numpy as np

import cv2

# 读取图片 (BGR 格式)
img = cv2.imread('photo.jpg')
img_color = cv2.imread('photo.jpg', cv2.IMREAD_COLOR)     # 彩色（默认）
img_gray = cv2.imread('photo.jpg', cv2.IMREAD_GRAYSCALE)  # 灰度

# 查看信息
print(f"形状: {img.shape}")       # (高度, 宽度, 通道数)
print(f"数据类型: {img.dtype}")   # uint8
print(f"像素总数: {img.size}")    # 高 × 宽 × 通道

# 显示图片
cv2.imshow('Image', img)
cv2.waitKey(0)       # 按任意键关闭
cv2.destroyAllWindows()

# 保存图片
cv2.imwrite('output.jpg', img, [cv2.IMWRITE_JPEG_QUALITY, 95])

# 读取单个像素 (BGR)
pixel = img[100, 200]    # [B, G, R] 值

# 修改像素
img[100, 200] = [255, 0, 0]  # 蓝色

# ROI（感兴趣区域）
roi = img[50:200, 100:300]   # (y1:y2, x1:x2)
img[50:200, 100:300] = 0     # 将该区域涂黑

# 使用掩码
mask = np.zeros(img.shape[:2], dtype=np.uint8)
mask[50:200, 100:300] = 255
result = cv2.bitwise_and(img, img, mask=mask)

import cv2
import numpy as np

img = cv2.imread('photo.jpg')

# 缩放
resized = cv2.resize(img, (400, 300))                    # 指定尺寸
scaled = cv2.resize(img, None, fx=0.5, fy=0.5)           # 按比例

# 旋转
h, w = img.shape[:2]
center = (w // 2, h // 2)
matrix = cv2.getRotationMatrix2D(center, angle=45, scale=1.0)
rotated = cv2.warpAffine(img, matrix, (w, h))

# 仿射变换
pts1 = np.float32([[0, 0], [w-1, 0], [0, h-1]])     # 原始3个点
pts2 = np.float32([[0, 0], [w-1, 0], [int(w*0.33), h-1]])  # 目标3个点
matrix = cv2.getAffineTransform(pts1, pts2)
warped = cv2.warpAffine(img, matrix, (w, h))

# 透视变换
pts1 = np.float32([[56, 65], [368, 52], [28, 387], [389, 390]])
pts2 = np.float32([[0, 0], [300, 0], [0, 300], [300, 300]])
matrix = cv2.getPerspectiveTransform(pts1, pts2)
perspective = cv2.warpPerspective(img, matrix, (300, 300))

img = cv2.imread('photo.jpg')

# 均值模糊
blur = cv2.blur(img, (5, 5))

# 高斯模糊（更自然的模糊效果）
gaussian = cv2.GaussianBlur(img, (5, 5), 0)

# 中值滤波（去除椒盐噪声效果最好）
median = cv2.medianBlur(img, 5)

# 双边滤波（保边去噪）
bilateral = cv2.bilateralFilter(img, 9, 75, 75)

img = cv2.imread('photo.jpg')
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

# Canny 边缘检测（最常用）
edges = cv2.Canny(gray, threshold1=50, threshold2=150)

# Sobel 算子
sobel_x = cv2.Sobel(gray, cv2.CV_64F, 1, 0, ksize=3)
sobel_y = cv2.Sobel(gray, cv2.CV_64F, 0, 1, ksize=3)
sobel = cv2.magnitude(sobel_x, sobel_y)

# Laplacian 算子
laplacian = cv2.Laplacian(gray, cv2.CV_64F)

# 边缘检测 + 二值化
_, binary = cv2.threshold(edges, 127, 255, cv2.THRESH_BINARY)

img = cv2.imread('photo.jpg')

# BGR → 灰度
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

# BGR → RGB（与PIL互转时需要）
rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)

# BGR → HSV（色相-饱和度-明度，颜色分割常用）
hsv = cv2.cvtColor(img, cv2.COLOR_BGR2HSV)

# BGR → LAB（感知均匀的颜色空间）
lab = cv2.cvtColor(img, cv2.COLOR_BGR2LAB)

# HSV 颜色分割示例
lower_red = np.array([0, 100, 100])
upper_red = np.array([10, 255, 255])
mask = cv2.inRange(hsv, lower_red, upper_red)
red_only = cv2.bitwise_and(img, img, mask=mask)

import cv2
import numpy as np

# 先二值化
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
_, binary = cv2.threshold(gray, 127, 255, cv2.THRESH_BINARY)

# 定义结构元素
kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (5, 5))
# 也可用圆形: cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (5, 5))

# 腐蚀 —— 收缩白色区域
eroded = cv2.erode(binary, kernel, iterations=1)

# 膨胀 —— 扩张白色区域
dilated = cv2.dilate(binary, kernel, iterations=1)

# 开运算 —— 先腐蚀后膨胀（去噪点）
opened = cv2.morphologyEx(binary, cv2.MORPH_OPEN, kernel)

# 闭运算 —— 先膨胀后腐蚀（填孔洞）
closed = cv2.morphologyEx(binary, cv2.MORPH_CLOSE, kernel)

# 形态学梯度 —— 膨胀 - 腐蚀（提取边缘）
gradient = cv2.morphologyEx(binary, cv2.MORPH_GRADIENT, kernel)

# 顶帽 —— 原图 - 开运算（提取小亮区域）
tophat = cv2.morphologyEx(binary, cv2.MORPH_TOPHAT, kernel)

# 黑帽 —— 闭运算 - 原图（提取小暗区域）
blackhat = cv2.morphologyEx(binary, cv2.MORPH_BLACKHAT, kernel)

import cv2

img = cv2.imread('shapes.jpg')
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
_, binary = cv2.threshold(gray, 127, 255, cv2.THRESH_BINARY)

# 查找轮廓
contours, hierarchy = cv2.findContours(binary, cv2.RETR_TREE,
                                         cv2.CHAIN_APPROX_SIMPLE)

# 绘制所有轮廓
result = img.copy()
cv2.drawContours(result, contours, -1, (0, 255, 0), 2)

# 分析每个轮廓
for i, contour in enumerate(contours):
    # 面积
    area = cv2.contourArea(contour)
    # 周长
    perimeter = cv2.arcLength(contour, True)
    # 多边形近似
    epsilon = 0.02 * perimeter
    approx = cv2.approxPolyDP(contour, epsilon, True)
    # 外接矩形
    x, y, w, h = cv2.boundingRect(contour)
    # 外接圆
    (cx, cy), radius = cv2.minEnclosingCircle(contour)

    print(f"轮廓 {i}: 面积={area:.0f}, 周长={perimeter:.0f}, "
          f"顶点数={len(approx)}, 类形={'矩形' if len(approx)==4 else '其他'}")

    # 绘制外接矩形和圆
    cv2.rectangle(result, (x, y), (x+w, y+h), (255, 0, 0), 2)
    cv2.circle(result, (int(cx), int(cy)), int(radius), (0, 0, 255), 2)

from PIL import Image
import cv2
import numpy as np

# Pillow → OpenCV
pil_img = Image.open('photo.jpg')
cv_img = np.array(pil_img)[:, :, ::-1]  # RGB → BGR

# OpenCV → Pillow
cv_img = cv2.imread('photo.jpg')
pil_img = Image.fromarray(cv2.cvtColor(cv_img, cv2.COLOR_BGR2RGB))