OpenCV---应用

模板匹配

import os
import matplotlib.pyplot as plt
import cv2

def opencv_compare(dst_target_dir, target, dst_template_dir, template):
    """
    获取 opencv操作后得到的 分值； 这里1为最大值；0.95为阈值，>0.95的为OK
    :param dst_target_dir: 须进入到目标目录使用;  这样可以防止中文路径
    :param dst_template_dir: 须进入到目标目录使用;  这样可以防止中文路径
    :param target:
    :param template:
    :return:
    """
    os.chdir(dst_target_dir)
    target_ = cv2.imread(target)
    # 读取模板图片
    os.chdir(dst_template_dir)
    template_ = cv2.imread(template)

    # 执行模板匹配，采用的匹配方式cv2.TM_CCOEFF_NORMED
    result = cv2.matchTemplate(target_, template_, cv2.TM_CCOEFF_NORMED)
    # 寻找矩阵（一维数组当做向量，用Mat定义）中的最大值和最小值的匹配结果及其位置
    min_val, max_val, min_loc, max_loc = cv2.minMaxLoc(result)
    # 打印出当前的值
    return max_val, max_loc

坐标值获取

def get_gray_xy(image_path) -> dict:
    """
    通过左键点击灰度图片，获取点击点的坐标
    可以在图上标出坐标值
    :param image_path:
    :return:
    """
    dic_xy = {}
    # 读取图片
    image = cv2.imread(image_path)
    # 图片转为灰度图
    gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
    dic_xy['pic_1x'] = 0
    dic_xy['pic_1y'] = 0

    def on_mouse(event, x, y, flags, param):
        if event == cv2.EVENT_LBUTTONDOWN:
            xy = "%d,%d" % (x, y)
            dic_xy['pic_1x'] = x
            dic_xy['pic_1y'] = y
            cv2.circle(image, (x, y), 2, (255, 0, 0), thickness=-1)
            cv2.putText(image, xy, (x, y), cv2.FONT_HERSHEY_PLAIN,
                        4.0, (0, 0, 255), thickness=3)
            cv2.imshow("image", image)
            print(type(x))
            print(dic_xy)

    cv2.namedWindow("image", cv2.WINDOW_NORMAL)
    cv2.setMouseCallback("image", on_mouse)
    cv2.imshow("image", image)

    cv2.waitKey(0)
    cv2.destroyAllWindows()
    return dic_xy

灰度值获取

def get_gray_value(image_path) -> int:
    """
    通过左键点击灰度图片，获取点击点的灰度值
    可以在图上标出灰度值
    :param image_path:
    :return:
    """
    # 读取图片
    image = cv2.imread(image_path)
    # 图片转为灰度图
    gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
    gray_value = 0

    def on_mouse(event, x, y, flags, param):
        if event == cv2.EVENT_LBUTTONDOWN:
            gray_value = int(gray[y, x])
            # gray_value_str = f'({x},{y}):{str(gray_value)}' # 带坐标
            gray_value_str = f'{str(gray_value)}'
            cv2.circle(image, (x, y), 2, (255, 0, 0), thickness=-1)
            cv2.putText(image, gray_value_str, (x, y), cv2.FONT_HERSHEY_PLAIN,
                        3.0, (0, 0, 255), thickness=3)
            cv2.imshow("image", image)

    cv2.namedWindow("image", cv2.WINDOW_NORMAL)
    cv2.setMouseCallback("image", on_mouse)
    cv2.imshow("image", image)

    cv2.waitKey(0)
    cv2.destroyAllWindows()
    return gray_value

特定像素点个数

def get_gray_white_num(image_path, start_gray, end_gray):
    """
    获取需要统计的像素数、总的像素数、比率
    :param end_gray: 开始的灰度值 如248
    :param start_gray: 结束的灰度值 如252
    :param image_path:
    :return:需要统计的像素数、总的像素数、比率
    """
    # 读取图片
    image = cv2.imread(image_path)
    # 图片转为灰度图
    gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
    # 总的像素值
    gray_size = gray.size
    # [gray]:图像；[0]:灰度；None:mask掩模图像；[256]：分为256份，一份1个的求和；[0,256]:像素范围
    hist = cv2.calcHist([gray], [0], None, [256], [0, 256])
    num = 0
    for i in range(start_gray, end_gray + 1):
        num = num + hist[i]
    return int(num), gray_size, round(int(num) / gray_size, 5)

阈值化后显示轮廓

def show_window_threshold(image_path, start_gray):
    """
    显示轮廓
    """
    im = cv2.imread(image_path)
    imgray = cv2.cvtColor(im, cv2.COLOR_BGR2GRAY)
    imgray_gauss = cv2.GaussianBlur(imgray, (3, 3), 0)  # 高斯滤波
    # 返回值： thresh:阈值化后的图像
    ret, thresh = cv2.threshold(imgray_gauss, start_gray, 255, 0)
    # 返回值：轮廓，轮廓层析结构
    # 轮廓是一个 Python列表，其中存储这图像中的所有轮廓。
    # 每一个轮廓都是一个 Numpy 数组，包含对象边界点（ x， y）的坐标
    contours, hierarchy = cv2.findContours(thresh, cv2.RETR_EXTERNAL, 
                                           cv2.CHAIN_APPROX_SIMPLE)
    # 画图 参数1：原始图件；参数2：轮廓；参数3：轮廓索引；参数45：颜色与厚度
    im = cv2.drawContours(im, contours, -1, (0, 255, 0), 3)
    cv2.namedWindow("image", cv2.WINDOW_NORMAL)
    cv2.imshow("image", im)
    cv2.waitKey()
    cv2.destroyAllWindows()

canny边缘检测后显示轮廓

def show_window_canny(image_path):
    """
    canny边缘检测后，画出边缘图
    """
    im = cv2.imread(image_path)
    imgray = cv2.cvtColor(im, cv2.COLOR_BGR2GRAY)
    # 高斯滤波，此处必须有(7,7)的核，不然会有噪点，影响最值点的判断
    imgray_gauss = cv2.GaussianBlur(imgray, (7, 7), 0)  
    canny = cv2.Canny(imgray_gauss, 50, 100)
    contours, hierarchy = cv2.findContours(canny, cv2.RETR_EXTERNAL,
                                           cv2.CHAIN_APPROX_SIMPLE)
    im = cv2.drawContours(im, contours, -1, (0, 255, 0), 3)  # 画图
    cv2.namedWindow("image", cv2.WINDOW_NORMAL)
    cv2.imshow("image", im)
    cv2.waitKey()
    cv2.destroyAllWindows()

只显示图片的G通道

def show_change(image_path):
    rgb_img = cv2.imread(image_path)  # 加载BGR彩色图像
    rgb_img[:, :, 0] = 0
    rgb_img[:, :, 2] = 0

    cv2.namedWindow('image', cv2.WINDOW_NORMAL)
    cv2.imshow('image', rgb_img)
    cv2.waitKey(0)
    cv2.destroyAllWindows()