python，opencv利用自适应阈值分割法实现微滴图像分割并计数

import cv2
import numpy as np

blockSize = 31
value = -1

#count = 0  #液滴总数
area = 0  #单个液滴面积
min_area = 40
max_area = 1500


#闭运算
def close(image):
    kernel = cv2.getStructuringElement(cv2.MORPH_ELLIPSE,(5,5))
    iClose = cv2.morphologyEx(image,cv2.MORPH_CLOSE,kernel)
    return iClose

#开运算
def open(image):
    kernel = cv2.getStructuringElement(cv2.MORPH_ELLIPSE,(16,16))
    iOpen = cv2.morphologyEx(image,cv2.MORPH_OPEN,kernel)
    return iOpen

#查找轮廓
def findConftours(srcImage,binary):
    contours,hierarchy = cv2.findContours(binary,cv2.RETR_TREE,cv2.CHAIN_APPROX_SIMPLE)
    cv2.drawContours(srcImage,contours,-1,(0,0,255),3)

#遍历所有的荧光区域,计数
def countAll(contours,image):
    global count
    count = 0
    for i in range(np.size(contours)):
        area = cv2.contourArea(contours[i])  #计算闭合轮廓面积
        if (area < min_area) or (area > max_area):
            continue
        else:
            count = count + 1
            (x,y),radius = cv2.minEnclosingCircle(contours[i])
            (x,y,radius) = np.int0((x,y,radius))
            cv2.circle(image,(x,y),radius,(0,0,255),2)
    return image,count


def cut():
    img = cv2.imread("E:/PythonWorkspace/yiingguang_image_process/images/1121.bmp")
    grayImage = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)
    img2 = img.copy()

    #ret, th1 = cv2.threshold(grayImage, 127, 255, cv2.THRESH_BINARY)
    # 第一个参数为原始图像矩阵，第二个参数为像素值上限，第三个是自适应方法（adaptive method）：
    #                                              -----cv2.ADAPTIVE_THRESH_MEAN_C:领域内均值
    #                                              -----cv2.ADAPTIVE_THRESH_GAUSSIAN_C:领域内像素点加权和，权重为一个高斯窗口
    # 第四个值的赋值方法：只有cv2.THRESH_BINARY和cv2.THRESH_BINARY_INV
    # 第五个Block size：设定领域大小（一个正方形的领域）
    # 第六个参数C，阈值等于均值或者加权值减去这个常数（为0相当于阈值，就是求得领域内均值或者加权值）
    # 这种方法理论上得到的效果更好，相当于在动态自适应的调整属于自己像素点的阈值，而不是整幅图都用一个阈值
    #th2 = cv2.adaptiveThreshold(grayImage, 255, cv2.ADAPTIVE_THRESH_MEAN_C, cv2.THRESH_BINARY, blockSize, value)
    #th3 = cv2.adaptiveThreshold(grayImage, 255, cv2.ADAPTIVE_THRESH_MEAN_C, cv2.THRESH_BINARY, blockSize, value)
    th4 = cv2.adaptiveThreshold(grayImage, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C, cv2.THRESH_BINARY, blockSize, value)
    #cv2.imshow('img', grayImage)
    #cv2.imshow('th1', th1)
    #cv2.imshow('th2', th2)
    #cv2.imshow('th3', th3)
    cv2.imshow('th4', th4)

    close_image = th4

    #去除噪声,形态学操作
    #闭运算
    iClose = close(close_image)
    cv2.imshow("close",iClose)
    #开运算后
    iOpen = open(iClose)
    cv2.imshow("close_and_open",iOpen)

    binary,contours,hirarchy = cv2.findContours(iOpen,cv2.RETR_EXTERNAL,cv2.CHAIN_APPROX_SIMPLE)
    print("所有的荧光区域："+format(np.size(contours)))
    #显示轮廓

    res = cv2.drawContours(img,contours,-1,(0,0,255),2)
    #tmp = np.zeros(img.shape,np.uint8)
    #res = cv2.drawContours(tmp, contours, -1, (0, 0, 255), 2)
    cv2.imshow("cut_res",res)
    cv2.imwrite("adaptiveThreshold_cut_res.bmp",res)

    #绘制最小外接圆
    res,count = countAll(contours,img2)
    cv2.imshow("cirle_res",res)
    cv2.imwrite("adaptiveThreshold_cirle_res.bmp",res)
    print("筛选后的荧光亮点数："+format(count))

    cv2.waitKey(0)
    cv2.destroyAllWindows()



if __name__ == '__main__':
    cut()

 

来源：CSDN

作者：Hu_AmI

链接：https://blog.csdn.net/qq_32295155/article/details/103633116