实战 | OpenCV两种不同方法实现粘连大米分割计数(步骤 + 代码)

Color Space

发布于 2024-01-18 14:33:28

2920

发布于 2024-01-18 14:33:28

文章被收录于专栏：OpenCV与AI深度学习OpenCV与AI深度学习

视觉/图像重磅干货，第一时间送达

导读

本文主要介绍基于OpenCV的两种不同方法实现粘连大米分割计数，并给详细步骤和代码。

背景介绍

测试图如下，图中有个别米粒相互粘连，本文主要演示如何使用OpenCV用两种不同方法将其分割并计数。

方法一：基于分水岭算法

基于分水岭算法分割步骤如下：

【1】高斯滤波 + 二值化 + 开运算

gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
gray = cv2.GaussianBlur(gray,(5,5),0)
ret, binary= cv2.threshold(gray, 115, 255, cv2.THRESH_BINARY)
 
kernel = np.ones((5, 5), np.uint8)
binary = cv2.morphologyEx(binary, cv2.MORPH_OPEN, kernel, iterations=1)
cv2.imshow('thres', binary)

【2】距离变换 + 提取前景

dist = cv2.distanceTransform(binary, cv2.DIST_L2, 3)
dist_out = cv2.normalize(dist, 0, 1.0, cv2.NORM_MINMAX)
cv2.imshow('distance-Transform', dist_out * 100)
ret, surface = cv2.threshold(dist_out, 0.35*dist_out.max(), 255, cv2.THRESH_BINARY)
cv2.imshow('surface', surface)
sure_fg = np.uint8(surface)# 转成8位整型
cv2.imshow('Sure foreground', sure_fg)

【3】标记位置区域

# 未知区域标记为0
markers[unknown == 255] = 0
kernel = np.ones((5, 5), np.uint8)
binary = cv2.morphologyEx(binary, cv2.MORPH_DILATE, kernel, iterations=1)
unknown = binary - sure_fg
cv2.imshow('unknown',unknown)

【4】分水岭算法分割

markers = cv2.watershed(img, markers=markers)
min_val, max_val, min_loc, max_loc = cv2.minMaxLoc(markers)

【5】轮廓查找和标记

contours,hierarchy = cv2.findContours(mask, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_NONE)
    for cnt in contours:
        M = cv2.moments(cnt)
        cx = int(M['m10']/M['m00'])
        cx = int(M['m10']/M['m00'])
        cy = int(M['m01']/M['m00'])#轮廓重心
        cv2.drawContours(img,contours,-1,colors[rd.randint(0,5)],2)
        cv2.drawMarker(img, (cx,cy),(0,255,0),1,8,2)

方法二：轮廓凸包缺陷方法

基于轮廓凸包缺陷分割步骤如下：

【1】高斯滤波 + 二值化 + 开运算

gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
gray = cv2.GaussianBlur(gray,(5,5),0)
ret, binary= cv2.threshold(gray, 115, 255, cv2.THRESH_BINARY)
 
kernel = np.ones((5, 5), np.uint8)
binary = cv2.morphologyEx(binary, cv2.MORPH_OPEN, kernel, iterations=1)
cv2.imshow('thres', binary)

【2】轮廓遍历 + 筛选轮廓含有凸包缺陷的轮廓

contours,hierarchy = cv2.findContours(binary, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_NONE)
for cnt in contours:
    hull = cv2.convexHull(cnt,returnPoints=False)#默认returnPoints=True
    defects = cv2.convexityDefects(cnt,hull)
    #print defects
    pt_list = []
    if defects is not None:
        flag = False
        for i in range(0,defects.shape[0]):
            s,e,f,d = defects[i,0]
            if d > 4500:
                flag = True

【3】将距离d最大的两个凸包缺陷点连起来，将二值图中对应的粘连区域分割开，红色圆标注为分割开的部分

    if len(pt_list) > 0:
        cv2.line(binary,pt_list[0],pt_list[1],0,2)  
cv2.imshow('binary2',binary)

【4】重新查找轮廓并标记结果

contours,hierarchy = cv2.findContours(binary, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_NONE)
for cnt in contours:
    try:
        M = cv2.moments(cnt)
        cx = int(M['m10']/M['m00'])
        cx = int(M['m10']/M['m00'])
        cy = int(M['m01']/M['m00'])#轮廓重心
         
        cv2.drawContours(img,cnt,-1,colors[rd.randint(0,5)],2)
        cv2.drawMarker(img, (cx,cy),(0,0,255),1,8,2)
    except:
        pass