Python使用OpenCV

怼烎@ 2022-05-23 12:23 186阅读 0赞

参考了：[https://blog.csdn.net/firemicrocosm/article/list/2?t=1][https_blog.csdn.net_firemicrocosm_article_list_2_t_1]  
在此表示感谢！  
基本图像操作函数：

## **图像的读取，保存，RGB分层** ##

# coding= utf-8 #or gbk 这样才能使用中文 
    
    import cv2  
    import numpy as np  
    
    **# 读取图像** 
    img = cv2.imread("lena.jpg", 1)   # 读取彩色图像 
    img_g = cv2.imread("lena.jpg", 0)  # 读取灰度图像 
    
    **# 复制图像** 
    img_copy = img.copy()  
    
    **# 获取图像的尺寸及通道数** 
    width, height, ch = img.shape  
    print(width, height, ch)  
    print img.size    # 打印像素数目 
    
    # 彩色通道的拆分与合并，速度较慢 
    b, g, r = cv2.split(img)  
    img = cv2.merge([b,g,r])  
    
    # 采用类似 matlab 中的下标索引方式，速度比较快 
    # 分离彩色通道，顺序是BGR,与 matlab的RGB 不同 
    B = img[:, :, 0]  
    G = img[:, :, 1]  
    R = img[:, :, 2]  
    
    # 调整RGB通道的顺序 
    img_RGB = img.copy()  
    img_RGB[:, :, 0] = R  
    img_RGB[:, :, 1] = G  
    img_RGB[:, :, 2] = B  
    
    # 采用类似 matlab 显示多个坐标轴的形式并行显示，比较方便 
    # 由于 matlab 是按照RGB顺序显示的，所以需要将原图像的顺序重调一下 
    plt.subplot(231), plt.imshow(img,'gray'), plt.title("img_BGR")  
    plt.subplot(232), plt.imshow(img_RGB, "gray"), plt.title("Img_RGB")  
    plt.subplot(233), plt.imshow(B, "gray"), plt.title("B")  
    plt.subplot(234), plt.imshow(G, "gray"), plt.title("G")  
    plt.subplot(235), plt.imshow(R, "gray"), plt.title("R")  
    plt.show()  
    
    # 利用NumPy 创建空的图像 
    emptyImage = np.zeros(img.shape, np.uint8)  #参数形状，位图深度
    
    # 保存图像 
    cv2.imwrite("lena_copy.jpg", img_copy)  
    
    # 颜色变换 
    img_gray = cv2.cvtColor(img, cv2.COLOR_RGB2GRAY)  
    
    # 大小尺度变换 
    img_resize = cv2.resize(img, None, fx=0.5, fy=0.5, interpolation=cv2.INTER_CUBIC)  
    
    cv2.namedWindow("lena_rgb")  
    cv2.namedWindow("lena_gray")  
    cv2.namedWindow("emptyImage")  
    cv2.namedWindow("gray")  
    cv2.namedWindow("resize")  
    
    cv2.imshow("lena_rgb", img)  
    cv2.imshow("lena_gray", img_g)  
    cv2.imshow("emptyImage", emptyImage)  
    cv2.imshow("gray", img_gray)  
    cv2.imshow("resize", img_resize)  
    
    cv2.waitKey(0)    #让窗口等待（不加则会一闪而过）
    cv2.destroyAllWindows()

## 图像滤波算子 ##

**平滑与锐化**

#coding:utf8
    import cv2
    import numpy as np
    from matplotlib import pyplot as plt
    
    
    ####################################
    # 将图像的BGR顺序纠正为RGB
    def bgr2rgb(src):
        img = src.copy()
        img[:, :, 0] = src[:, :, 2]
        img[:, :, 2] = src[:, :, 0]
        return img
    
    
    ######################################
    
    # 图像平滑
    
    img = cv2.imread("1.jpg")
    
    # 创建滤波器
    k = 9
    kernel = np.ones((k, k), np.float32) / k ** 2
    print kernel
    
    # 使用自滤波器进行滤波
        # cv.Filter2D(src, dst, kernel, anchor=(-1, -1))
        # ddepth –desired depth of the destination image;
        # if it is negative, it will be the same as src.depth();
        # the following combinations of src.depth() and ddepth are supported:
        # src.depth() = CV_8U, ddepth = -1/CV_16S/CV_32F/CV_64F
        # src.depth() = CV_16U/CV_16S, ddepth = -1/CV_32F/CV_64F
        # src.depth() = CV_32F, ddepth = -1/CV_32F/CV_64F
        # src.depth() = CV_64F, ddepth = -1/CV_64F
        # when ddepth = -1, the output image will have the same depth as the source.
    img_average = cv2.filter2D(img, -1, kernel)
    
    # 使用blur()函数进行均值滤波
    img_blur = cv2.blur(img, (k, k))
    
    # 高斯平滑
    img_gaussian = cv2.GaussianBlur(img, (k, k), 3)
    # 也可以自己创建一个高斯核
    # kernel_g = cv2.getGaussianKernel(k, 5)
    # img_gaussian = cv2.filter2D(img,-1,kernel_g)
    # print kernel_g.shape
    
    # 中值滤波
    # img_median = cv2.medianBlur(img_noise,k)
    
    # plt.subplot(121),plt.imshow(img_noise,'gray'),plt.title("噪声图像")
    # plt.subplot(122),plt.imshow(img_median,'gray'),plt.title("中值滤波后的图像")
    # plt.show()
    
    # 双边滤波
    # cv2.bilateralFilter(src, d, sigmaColor, sigmaSpace)
    # d – Diameter of each pixel neighborhood that is used during filtering.
    # If it is non-positive, it is computed from sigmaSpace
    # 9 邻域直径，两个75 分别是空间高斯函数标准差，灰度值相似性高斯函数标准差
    img_new = cv2.imread("Airplane.jpg")
    img_bf = cv2.bilateralFilter(img_new, 9, 75, 75)
    
    # 平滑滤波结果
    plt.subplot(221), plt.imshow(bgr2rgb(img), "gray"), plt.title("Original")
    plt.xticks([]), plt.yticks([])  # 去掉坐标轴刻度
    plt.subplot(222), plt.imshow(bgr2rgb(img_average), "gray"), plt.title("average_filtering")
    plt.xticks([]), plt.yticks([])
    plt.subplot(223), plt.imshow(bgr2rgb(img_blur), "gray"), plt.title("blur_filtering")
    plt.xticks([]), plt.yticks([])
    plt.subplot(224), plt.imshow(bgr2rgb(img_gaussian), "gray"), plt.title("gaussian_filtering")
    plt.xticks([]), plt.yticks([])
    plt.show()
    
    # 双边滤波结果
    plt.subplot(121), plt.imshow(bgr2rgb(img_new), "gray"), plt.title("Original")
    plt.xticks([]), plt.yticks([])
    plt.subplot(122), plt.imshow(bgr2rgb(img_bf), "gray"), plt.title("bilateralFilter")
    plt.xticks([]), plt.yticks([])
    plt.show()
    
    ##锐化
    img_gray = cv2.cvtColor(img, cv2.COLOR_RGB2GRAY)      #转化为灰度图
    
    # sobel 算子 
    # cv2.Sobel(src,ddepth,dx,dy,dst=None,ksize,scale,delta) 
    # dx = 1 对x方向求梯度 
    # dy = 1 对y方向求梯度 
    img_sobel_x = cv2.Sobel(img_gray, cv2.CV_64F, 1, 0, ksize=3)  
    img_sobel_y = cv2.Sobel(img_gray, cv2.CV_64F, 0, 1, ksize=3)  
    
    # Laplace 算子 
    img_laplace = cv2.Laplacian(img_gray, cv2.CV_64F, ksize=3)  
    
    # Canny 算子 
    img_canny = cv2.Canny(img_gray, 100 , 150)  
    
    plt.subplot(231), plt.imshow(img_gray, "gray"), plt.title("Original")  
    plt.subplot(232), plt.imshow(img_sobel_x, "gray"), plt.title("Sobel_x")  
    plt.subplot(233), plt.imshow(img_sobel_y, "gray"), plt.title("Sobel_y")  
    plt.subplot(234), plt.imshow(img_laplace,  "gray"), plt.title("Laplace")  
    plt.subplot(235), plt.imshow(img_canny, "gray"), plt.title("Canny")  
    plt.show()

![这里写图片描述][70]

## 直方图 ##

# coding:utf-8 
    
    import cv2  
    import numpy as np  
    from matplotlib import pyplot as plt  
    
    file = "E:\python\Python Project\opencv_showimage\images\lena.jpg"  
    img = cv2.imread(file, 0)  
    # OpenCV 函数 
    hist = cv2.calcHist([img],[0],None,[256],[0,256])  
    plt.plot(hist, color='r')  
    plt.xlim([0,256])  
    plt.show()  
    
    # Numpy 函数 
    # ravel 是将矩阵拉成列向量 
    plt.hist(img.ravel(),256,[0,256])  
    plt.xlim([0,256])  
    plt.show()  
    
    # 彩色图像直方图 
    img = cv2.imread(file,1)  
    color = ('b','g','r')  
    
    for i,col in enumerate(color):  
        hist = cv2.calcHist([img],[i],None,[256],[0,256])  
        plt.plot(hist,color=col)  
        plt.xlim([0,256])  
    plt.show()  
    
    # 设置掩膜，只统计掩膜区域的直方图 
    # 创造一个掩膜 
    img = cv2.imread(file,0)  
    mask = np.zeros(img.shape[:2], np.uint8)  
    mask[100:300,100:400] = 255  
    masked_img = cv2.bitwise_and(img,img,mask=mask)  
    
    hist_full = cv2.calcHist([img],[0],None,[256],[0,256])  
    hist_mask = cv2.calcHist([img],[0],mask,[256],[0,256])  
    
    plt.subplot(221),plt.imshow(img,"gray")  
    plt.subplot(222),plt.imshow(mask,"gray")  
    plt.subplot(223),plt.imshow(masked_img,"gray")  
    plt.subplot(224),plt.plot(hist_full,color='r'),plt.plot(hist_mask,color='g'),plt.xlim([0,256])  
    plt.show()  
    
    # 直方图均衡化 
    img = cv2.imread(file, 0)  
    equ = cv2.equalizeHist(img)  
    res = np.hstack((img, equ))   # 将图像拼在一起 
    
    cv2.namedWindow("equ")  
    cv2.imshow("equ", res)  
    cv2.waitKey(0)  
    cv2.destroyAllWindows()  
    
    # CLAHE 有限对比适应性直方图均衡化,效果更好 
    clahe = cv2.createCLAHE(clipLimit=2.0, tileGridSize=(8,8))  
    cl1 = clahe.apply(img)  
    res1 = np.hstack((equ, cl1))   # 将图像拼在一起 
    
    cv2.imshow("CLAHE", res1)  
    cv2.waitKey(0)  
    cv2.destroyAllWindows()  
    
    # 2D 直方图 
    img = cv2.imread(file, 1)  
    img_hsv = cv2.cvtColor(img, cv2.COLOR_BGR2HSV)  
    hist = cv2.calcHist([img_hsv],[0,1],None,[180,256],[0,180,0,256])  
    plt.imshow(hist, interpolation='nearest'), plt.xlabel("S"), plt.ylabel("H")  
    plt.show()

RGB颜色直方图：  
![这里写图片描述][70 1]

直方图均衡化：增强对比度  
![这里写图片描述][70 2]

有限对比适应性直方图均衡化,效果更好：  
![这里写图片描述][70 3]

[https_blog.csdn.net_firemicrocosm_article_list_2_t_1]: https://blog.csdn.net/firemicrocosm/article/list/2?t=1
[70]: /images/20220523/bbb83fb95c1b445c82748f4e5bdae303.png
[70 1]: /images/20220523/2b4647ac1e974e11a89d234af89d7a21.png
[70 2]: /images/20220523/41dd2d5739124acaa5d3138588fd09d1.png
[70 3]: /images/20220523/c8e50561b1874573bd2dcc7a53393954.png