python opencv 总结

1. 图像的几何变换

imgobj = cv2.imread('pho.png') #读取图像

cv2.namedWindow("image") #创建窗口并显示的是图像类型

cv2.imshow("image",imgobj)

cv2.waitKey(0)        #等待事件触发，参数0表示永久等待

cv2.destroyAllWindows()   #释放窗口

# resize

res = cv2.resize(imgobj,None,fx=0.5, fy=0.5, interpolation = cv2.INTER_CUBIC)

cv2.imshow("image",res)

#warpAffine   仿射

rows,cols = res.shape[0:2]

M = np.float32([[1,0,100],[0,1,50]])

dst = cv2.warpAffine(res,M,(cols,rows))

cv2.imshow('img',dst)

#getRotationMatrix2D  旋转

M = cv2.getRotationMatrix2D((cols/2,rows/2),90,1)#长宽减半 逆时针90度 缩放1倍

dst2 = cv2.warpAffine(res,M,(cols,rows))

cv2.imshow('img',dst2)

#flip 翻转

dst3=cv2.flip(res,1)#0延x轴翻转  任意正数延y轴翻转  任意负数延x/y轴同时翻转

cv2.imshow('img',dst3)

#二值化

gray = cv2.cvtColor(res, cv2.COLOR_BGR2GRAY)

ret, binary = cv2.threshold(gray, 0, 255, cv2.THRESH_BINARY_INV | cv2.THRESH_OTSU)

cv2.imshow('img',binary)

cv2.waitKey(0)

cv2.destroyAllWindows()

 

 

1. 霍夫变换  圆检测

def circle_detect(img):

    gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

    trans = cv2.GaussianBlur(gray, (9, 9), 0)

    ret_1, thresh_1 = cv2.threshold(trans, 40, 255, cv2.THRESH_BINARY)

   circles_2 = cv2.HoughCircles(thresh_1, cv2.HOUGH_GRADIENT, 1, 10, param1=1000,        param2=20)

    local_2 = circles_2[0][0]

    x_2 = int(local_2[0])

    y_2 = int(local_2[1])

    r_2 = int(local_2[2])

    img = cv2.circle(img, (x_2, y_2), r_2, (0, 255, 0), 2)

    n = 10

    img = cv2.rectangle(img, (x_2 - r_2 - n, y_2 - r_2 - n), (x_2 + r_2 + n, y_2 + r_2 + n), (255,             255, 0), 3)

    return img

 

if __name__ == '__main__':

    img = cv2.imread('rec_cir.jpg')

    # print(img.shape)

    circle_detect(img)

    # cv2.namedWindow('addImage')

    cv2.imshow('img_add', img)

    cv2.waitKey(0)

cv2.destroyAllWindows()

 

1. 轮廓检测   形状检测

需要进一步预处理  获取更好的特征

import cv2 as cv

import numpy as np

class ShapeAnalysis:

    def __init__(self):

        self.shapes = {'triangle': 0, 'rectangle': 0, 'polygons': 0, 'circles': 0}

    def analysis(self, frame):

        h, w, ch = frame.shape

        result = np.zeros((h, w, ch), dtype=np.uint8)

        # 二值化图像

        print("start to detect lines...\n")

        gray = cv.cvtColor(frame, cv.COLOR_BGR2GRAY)

        trans = cv.GaussianBlur(gray, (9, 9), 0)

        ret_1, thresh_1 = cv.threshold(trans, 40, 255, cv.THRESH_BINARY)

        # ret, binary = cv.threshold(gray, 0, 255, cv.THRESH_BINARY_INV | cv.THRESH_OTSU)

        cv.imshow("input image", frame)

        contours, hierarchy = cv.findContours(thresh_1, cv.RETR_EXTERNAL, cv.CHAIN_APPROX_SIMPLE)

        for cnt in range(len(contours)):

            # 提取与绘制轮廓

            cv.drawContours(result, contours, cnt, (0, 255, 0), 2)

            # 轮廓逼近

            epsilon = 0.01 * cv.arcLength(contours[cnt], True)

            approx = cv.approxPolyDP(contours[cnt], epsilon, True)

 

            # 分析几何形状

            corners = len(approx)

            shape_type = ""

            if corners == 3:

                count = self.shapes['triangle']

                count = count+1

                self.shapes['triangle'] = count

                shape_type = "三角形"

            if corners == 4:

                count = self.shapes['rectangle']

                count = count + 1

                self.shapes['rectangle'] = count

                shape_type = "矩形"

            if corners >= 10:

                count = self.shapes['circles']

                count = count + 1

                self.shapes['circles'] = count

                shape_type = "圆形"

            if 4 < corners < 10:

                count = self.shapes['polygons']

                count = count + 1

                self.shapes['polygons'] = count

                shape_type = "多边形"

            # 求解中心位置

            mm = cv.moments(contours[cnt])

            cx = int(mm['m10'] / mm['m00'])

            cy = int(mm['m01'] / mm['m00'])

            cv.circle(result, (cx, cy), 3, (0, 0, 255), -1)

            # 颜色分析

            color = frame[cy][cx]

            color_str = "(" + str(color[0]) + ", " + str(color[1]) + ", " + str(color[2]) + ")"

            # 计算面积与周长

            p = cv.arcLength(contours[cnt], True)

            area = cv.contourArea(contours[cnt])

            print("周长: %.3f, 面积: %.3f 颜色: %s 形状: %s "% (p, area, color_str, shape_type))

        cv.imshow("Analysis Result", self.draw_text_info(result))

        cv.imwrite("test-result.png", self.draw_text_info(result))

        return self.shapes

 

    def draw_text_info(self, image):

        c1 = self.shapes['triangle']

        c2 = self.shapes['rectangle']

        c3 = self.shapes['polygons']

        c4 = self.shapes['circles']

        cv.putText(image, "triangle: "+str(c1), (10, 20), cv.FONT_HERSHEY_PLAIN, 1.2, (255, 0, 0), 1)

        cv.putText(image, "rectangle: " + str(c2), (10, 40), cv.FONT_HERSHEY_PLAIN, 1.2, (255, 0, 0), 1)

        cv.putText(image, "polygons: " + str(c3), (10, 60), cv.FONT_HERSHEY_PLAIN, 1.2, (255, 0, 0), 1)

        cv.putText(image, "circles: " + str(c4), (10, 80), cv.FONT_HERSHEY_PLAIN, 1.2, (255, 0, 0), 1)

        return image

 

if __name__ == "__main__":

    src = cv.imread("rec_cir.jpg")

    ld = ShapeAnalysis()

    ld.analysis(src)

    cv.waitKey(0)

cv.destroyAllWindows()

 

1. 图像分割

#去噪求梯度

    gray = cv2.cvtColor(img_cv2, cv2.COLOR_BGR2GRAY)

    trans = cv2.GaussianBlur(gray, (9, 9), 0)

    gradX = cv2.Sobel(gray, ddepth=cv2.CV_32F, dx=1, dy=0)

    gradY = cv2.Sobel(gray, ddepth=cv2.CV_32F, dx=0, dy=1)

    gradient = cv2.subtract(gradX, gradY)

    gradient = cv2.convertScaleAbs(gradient)

 

    #去噪 二值化

    blurred = cv2.GaussianBlur(gradient, (9, 9), 0)

    (_, thresh) = cv2.threshold(blurred, 90, 255, cv2.THRESH_BINARY)

    #形态 腐蚀膨胀

    kernel = cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (25, 25))

    closed = cv2.morphologyEx(thresh, cv2.MORPH_CLOSE, kernel)

    closed = cv2.erode(closed, None, iterations=4)

    closed = cv2.dilate(closed, None, iterations=4)

 

 

    ret_1, binary = cv2.threshold(thresh, 90, 255, cv2.THRESH_BINARY)

    # ret, binary = cv2.threshold(gray, 0, 255, cv2.THRESH_BINARY_INV | cv2.THRESH_OTSU)

    contours, hierarchy= cv2.findContours(binary, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)

 

    c = sorted(contours, key=cv2.contourArea, reverse=True)[0]

    # compute the rotated bounding box of the largest contour

    rect = cv2.minAreaRect(c)

    box = np.int0(cv2.boxPoints(rect))

    # draw a bounding box arounded the detected barcode and display the image

    # draw_img = cv2.drawContours(img_cv2, [box], -1, (0, 0, 255), 3)

    # cv2.imshow("draw_img", draw_img)

           cv2.drawContours(img_cv2, [box], -1, (0, 0, 255), 3)

    for cnt in range(len(contours)):

        # 提取与绘制轮廓

        cv2.drawContours(img_cv2, contours, cnt, (0, 255, 0), 3)

      

但目标提取效果很好，当图片出现多目标分割且背景复杂或有噪声情况下，无法准确分割出目标图像

 

1. createBackgroundSubtractorMOG2背景去除

import numpy as np

import cv2

cap = cv2.VideoCapture(0)

fgbg = cv2.createBackgroundSubtractorMOG2(detectShadows = False)

# fgbg = cv2.createBackgroundSubtractorKNN()

while(cap.isOpened()):

    ret, frame = cap.read()

    fgmask = fgbg.apply(frame)

    cv2.imshow('frame',fgmask)

    k = cv2.waitKey(30) & 0xff

    if k == 27:

        break

cap.release()

cv2.destroyAllWindows()