【OpenCV】Mango的OpenCV学习笔记【二】

最新推荐文章于 2024-08-11 15:51:22 发布

MangoloD

最新推荐文章于 2024-08-11 15:51:22 发布

阅读量247

点赞数

分类专栏： OpenCV 文章标签： opencv python

本文链接：https://blog.csdn.net/weixin_42166222/article/details/109688892

版权

OpenCV 专栏收录该内容

6 篇文章 1 订阅

订阅专栏

本文主要参考自 OpenCV官方文档

一、图像的基本操作

访问、修改像素值

import cv2

img = cv2.imread('../../Resources/messi15.jpg')
px = img[100, 100]
print(px)

# 仅访问单一通道像素值
blue = img[100, 100, 0]
print(blue)

# 修改像素值
img[100, 100] = [255, 255, 255]
print(img[100, 100])

# 更好的像素访问和编辑方法，返回标量
# 访问 RED 值
print(img.item(10, 10, 2))

# 修改RED值
img.itemset((10, 10, 2), 100)
print(img.item(10, 10, 2))

访问图像属性

# 图像形状
print(img.shape)

# 图像像素值总数
print(img.size)

# 图像数据类型
print(img.dtype)

图像感兴趣区域ROI
使用Numpy索引获得ROI，选择球并将其复制到图像中的另一个区域：

import cv2

img = cv2.imread('../../Resources/messi15.jpg')
ball = img[280:340, 330:390]
img[273:333, 100:160] = ball

cv2.imshow('messi', img)
cv2.waitKey(0)
cv2.destroyAllWindows()

效果图如下：在这里插入图片描述

为图像设置边框（了解即可）
如果要在图像周围创建边框（如相框），则可以使用 cv.copyMakeBorder() 。
其参数如下：
src：输入图像

top，bottom，left，right：边界宽度（以相应方向上的像素数为单位）

borderType：定义要添加哪种边框的标志。它可以是以下类型：

cv.BORDER_CONSTANT：添加恒定的彩色边框。该值应作为下一个参数给出。

cv.BORDER_REFLECT：边框将是边框元素的镜像，如下所示： fedcba | abcdefgh | hgfedcb

cv.BORDER_REFLECT_101 或 cv.BORDER_DEFAULT 与上述相同，但略有变化，例如： gfedcb | abcdefgh | gfedcba

cv.BORDER_REPLICATE：最后一个元素被复制，像这样： aaaaaa | abcdefgh | hhhhhhh

cv.BORDER_WRAP：难以解释，它看起来像这样： cdefgh | abcdefgh | abcdefg

value：边框的颜色，如果边框类型为 cv.BORDER_CONSTANT

import cv2
from matplotlib import pyplot as plt

BLUE = [255, 0, 0]
img1 = cv2.imread('../../Resources/opencv-logo.jpg')
replicate = cv2.copyMakeBorder(img1, 10, 10, 10, 10, cv2.BORDER_REPLICATE)
reflect = cv2.copyMakeBorder(img1, 10, 10, 10, 10, cv2.BORDER_REFLECT)
reflect101 = cv2.copyMakeBorder(img1, 10, 10, 10, 10, cv2.BORDER_REFLECT_101)
wrap = cv2.copyMakeBorder(img1, 10, 10, 10, 10, cv2.BORDER_WRAP)
constant = cv2.copyMakeBorder(img1, 10, 10, 10, 10, cv2.BORDER_CONSTANT, value=BLUE)
print(img1, reflect)
plt.subplot(231), plt.imshow(img1, 'gray'), plt.title('ORIGINAL')
plt.subplot(232), plt.imshow(replicate, 'gray'), plt.title('REPLICATE')
plt.subplot(233), plt.imshow(reflect, 'gray'), plt.title('REFLECT')
plt.subplot(234), plt.imshow(reflect101, 'gray'), plt.title('REFLECT_101')
plt.subplot(235), plt.imshow(wrap, 'gray'), plt.title('WRAP')
plt.subplot(236), plt.imshow(constant, 'gray'), plt.title('CONSTANT')
plt.show()

效果图如下（图像是采用matplotlib显示的，因此红色和蓝色通道将互换）：在这里插入图片描述

二、图像运算

图像加法
注意： OpenCV 加法和 Numpy 加法之间有区别
① OpenCV 加法是饱和运算
② 而 Numpy 加法是模运算

import cv2

img1 = cv2.imread('../../Resources/3D-Matplotlib.png')
img2 = cv2.imread('../../Resources/mainsvmimage.png')

add = img1 + img2  # 模运算，250+10 = 260 % 256 = 4
cv_add = cv2.add(img1, img2)  # 像素点值相加，所以图片变白，50+10 = 260 => 255

cv2.imshow('add', add)
cv2.imshow('cv_add', cv_add)
cv2.waitKey(0)
cv2.destroyAllWindows()

图像融合
公式： $d s t = s r c 1 * α + s r c 2 * β + γ$
```
import cv2

img1 = cv2.imread('./Image/3D-Matplotlib.png')
img2 = cv2.imread('./Image/mainsvmimage.png')
weighted = cv2.addWeighted(img1, 0.6, img2, 0.4, 0)  # 设置权重进行叠加
cv2.imshow('weighted', weighted)
cv2.waitKey(0)
cv2.destroyAllWindows()
```
效果图如下

按位运算
① 或运算，求并集，加法：cv2.bitwise_or(img1, img2)
② 与运算，求交集，重叠部位： cv2.bitwise_and(img1, img2)
③ 非运算，求空集，取反色：cv2.bitwise_not(img1)
④ 异或运算，相同为0，不同为1：
cv2.bitwise_xor(img1, img1)，cv2.bitwise_xor(img1, img2)

img1 = cv2.imread('../../Resources/3D-Matplotlib.png')
img2 = cv2.imread('../../Resources/mainlogo.png')

rows, cols, channels = img2.shape
roi = img1[0:rows, 0:cols]

img2_gray = cv2.cvtColor(img2, cv2.COLOR_BGR2GRAY)
ret, mask = cv2.threshold(img2_gray, 220, 255, cv2.THRESH_BINARY_INV)

mask_inv = cv2.bitwise_not(mask)
img1_bg = cv2.bitwise_and(roi, roi, mask=mask_inv)
img2_fg = cv2.bitwise_and(img2, img2, mask=mask)

dst = cv2.add(img1_bg, img2_fg)
img1[0:rows, 0:cols] = dst

# cv2.imshow('img_', img2_gray)
# cv2.imshow('img_1', mask)
# cv2.imshow('img_2', mask_inv)
# cv2.imshow('img_3', img1_bg)
# cv2.imshow('img_4', img2_fg)
# cv2.imshow('mask', mask)
cv2.imshow('new_img', img1)
cv2.waitKey(0)
cv2.destroyAllWindows()

效果图如下

三、视频练习

要求：向视频中添加动态字幕、矩形…

import cv2
import random
import numpy as np
import pandas as pd
from PIL import Image, ImageDraw, ImageFont
import time


# 颜色
def rand_color():
    font_r = random.randint(128, 255)
    font_g = random.randint(128, 255)
    font_b = random.randint(128, 255)
    return font_r, font_g, font_b


# 矩形坐标读取
def rectangle_point(index):
    data = pd.read_table("coordinate.txt", sep=",", header=None)
    point_x, point_y = [], []
    for i_ in range(len(data)):
        point_1 = (data[0][i_], data[1][i_])
        point_2 = (data[2][i_], data[3][i_])
        point_x.append(point_1)
        point_y.append(point_2)
    return point_x[index], point_y[index]


# 画矩形框
def draw_rectangle(frame_, point1, point2):
    cv2.rectangle(frame_, point1, point2, rand_color(), random.randint(1, 2))


# 中文读取
def chinese_text():
    data_1 = pd.read_table("text.txt", header=None)
    text_list = []
    for i_ in range(len(data_1)):
        text_list.append(data_1[0][i_])
    return text_list


# 绘制中文字体
def paint_chinese(frame_, chinese, pos, color):
    img_ = Image.fromarray(cv2.cvtColor(frame_, cv2.COLOR_BGR2RGB))
    draw = ImageDraw.Draw(img_)

    font = ImageFont.truetype(r'./msyh.ttc', 20, encoding='utf-8')
    fill_color = color  # (255,0,0)
    position = pos  # (100,100)

    draw.text(position, chinese, font=font, fill=fill_color)
    image_ = cv2.cvtColor(np.array(img_), cv2.COLOR_RGB2BGR)
    return image_


# 视频保存
def save_video(cap_):
    width_ = int(cap_.get(3))
    height_ = int(cap_.get(4))
    fps_ = cap.get(cv2.CAP_PROP_FPS)
    fourcc_ = cv2.VideoWriter_fourcc(*'mp4v')
    new_video = cv2.VideoWriter(r'./new_video.mp4', fourcc_, fps_, (width_, height_))
    return new_video


if __name__ == '__main__':
    start = time.time()
    path = r'../../Resources/cat.mp4'
    cap = cv2.VideoCapture(path)
    new_video_ = save_video(cap)

    fps = cap.get(cv2.CAP_PROP_FPS)
    total_fps = cap.get(cv2.CAP_PROP_FRAME_COUNT)

    current_frame = 0
    while True:
        ret, frame = cap.read()
        if cv2.waitKey(int(np.ceil(total_fps / fps)) // 2) & 0xFF == ord('q'):
            break
        elif ret is False:
            break

        for i in range(len(chinese_text())):
            start_frame = i * (total_fps / 5)
            end_frame = (i + 1) * (total_fps / 5)
            # print(current_frame, start_frame, end_frame, rectangle_point(i)[0], rectangle_point(i)[1])
            if start_frame <= current_frame <= end_frame:
                draw_rectangle(frame, rectangle_point(i)[0], rectangle_point(i)[1])
                new_img = paint_chinese(frame, chinese_text()[i], (200, 750), rand_color())

                new_video_.write(new_img)
                cv2.imshow('cat', new_img)

            if current_frame > total_fps:
                break

        current_frame += 1

    cap.release()
    new_video_.release()
    cv2.destroyAllWindows()
    end = time.time()
    print(end - start)