import numpy as np
import cv2

def update(val = 0):
    # disparity range is tuned for 'aloe' image pair
    stereo.setBlockSize(cv2.getTrackbarPos('window_size','disparity'))
    stereo.setUniquenessRatio(cv2.getTrackbarPos('uniquenessRatio','disparity'))
    stereo.setSpeckleWindowSize(cv2.getTrackbarPos('speckleWindowsize','disparity'))
    stereo.setSpeckleRange(cv2.getTrackbarPos('speckleRange','disparity'))
    stereo.setDisp12MaxDiff(cv2.getTrackbarPos('disp12MaxDiff','disparity'))

    print('computing disparity...')

    disp = stereo.compute(imgL, imgR).astype(np.float32) / 16.0

    # cv2.imshow('left', imgL)
    cv2.imshow('disparity', (disp-min_disp) / num_disp)

if __name__ == '__main__':
    window_size = 5
    min_disp = 16
    num_disp = 192 - min_disp
    blockSize = window_size
    uniquenessRatio = 1
    speckleRange = 3
    speckleWindowSize = 3
    disp12MaxDiff = 200
    P1 = 600
    P2 = 2400
    # 加载两幅图
    imgL = cv2.imread('tsukuba_right.jpg')
    imgR = cv2.imread('tsukuba_left.jpg')
    cv2.namedWindow('disparity')
    cv2.createTrackbar('speckleRange','disparity',speckleRange,50,update)
    cv2.createTrackbar('window_size','disparity',window_size,21,update)
    cv2.createTrackbar('speckleWindowSize','disparity',speckleWindowSize,200,update)
    cv2.createTrackbar('uniquenessRatio','disparity',uniquenessRatio,50,update)
    cv2.createTrackbar('disp12MaxDiff','disparity',disp12MaxDiff,250,update)

    # 创建一个StereoSGBM实例，是一种计算视图差的算法
    # 并创建几个跟踪条来调整算法参数，然后调用update函数
    # update函数将跟踪条的值传给StereoSGBM实例
    # StereoSGBM是semiglobal block matching 的缩写
    stereo = cv2.StereoSGBM_create(
        minDisparity = min_disp,
        numDisparities = num_disp,
        blockSize = window_size,
        uniquenessRatio = uniquenessRatio,
        speckleRange = speckleRange,
        speckleWindowSize = speckleWindowSize,
        disp12MaxDiff = disp12MaxDiff,
        P1 = P1,
        P2 = P2
    )
    update()
    cv2.waitKey()

cv2.grabCut(img, mask, rect, bgdModel, fgdModel, iterCount[, mode]) -> mask, bgdModel, fgdModel

import numpy as np
import cv2
from matplotlib import pyplot as plt

#使用分水岭和GrabCut算法进行物体分割
img = cv2.imread('small.jpg')

# img.shape=(1039, 690, 3)
# img.shape[0:2]=(1039, 690)
mask = np.zeros(img.shape[:2],np.uint8)

# 背景色bgdModel，前景色fgdModel
bgdModel = np.zeros((1,65),np.float64)
fgdModel = np.zeros((1,65),np.float64)

# 感兴趣区域ROI的x，y，宽度，高度
rect = (100,1,500,1000)

# 获得返回值mask、bgdModel、fgdModel。
# 目标图像、掩码、感兴趣区域，背景、前景、算法迭代次数、操作模式
cv2.grabCut(img,mask,rect,bgdModel,fgdModel,5,cv2.GC_INIT_WITH_RECT)

# 经过图像分割法grabCut处理之后，
# print(set(mask.ravel())) -> {0,2,3}
# mask的掩码元素{0}->{0,2,3}
# where(condition,x,y)，condition为array_like或bool
# 真yield x，假yield y
# mask==0背景、==1前景、==2可能的背景、==3可能的前景
# 当为背景/可能是背景时赋0，当为前景/可能背景赋1
mask2 = np.where((mask==2)|(mask==0),0,1).astype('uint8')

# mask2.shape=(1039,690),
# img.shape=(1039,690,3)
# 两者乘积则报错：
# 操作数无法与形状一起广播
# ValueError: operands could not be broadcast together with shapes (1039,690,3) (1039,690)
# 为了保持数形一致，增加np.newaxis
# mask2[:,:,np.newaxis].shape=(1039,690,1)
# 这样当行列值不相等时可进行广播计算
# 经过计算后，将背景色赋值为0，即为黑色
img = img*mask2[:,:,np.newaxis]

# subplot(121)创建1行2列，当前位置为1
plt.subplot(121), plt.imshow(img)
plt.title("grabcut"), plt.xticks([]), plt.yticks([])
# subplot(122)当前位置为2
plt.subplot(122), plt.imshow(cv2.cvtColor(cv2.imread('small.jpg'), cv2.COLOR_BGR2RGB))
plt.title("original"), plt.xticks([]), plt.yticks([])
plt.show()

cv2.watershed(image, markers) -> markers

import numpy as np
import cv2
from matplotlib import pyplot as plt

#使用分水岭算法进行图像分割
img = cv2.imread('timg.jpg')
gray = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)

#将颜色转为灰度后，可为图像设一个阈值，将图像分为两部分：黑色部分和白色部分
ret, thresh = cv2.threshold(gray,0,255,cv2.THRESH_BINARY_INV+cv2.THRESH_OTSU)

# noise removal 噪声去除，morphologyEx是一种对图像进行膨胀之后再进行腐蚀的操作
kernel = np.ones((3,3),np.uint8)
opening = cv2.morphologyEx(thresh,cv2.MORPH_OPEN,kernel, iterations = 2)

# sure background area 确定背景区域，图像进行膨胀操作
sure_bg = cv2.dilate(opening,kernel,iterations=3)

# Finding sure foreground area，通过distanceTransform来获取确定的前景区域
dist_transform = cv2.distanceTransform(opening,cv2.DIST_L2,5)
ret, sure_fg = cv2.threshold(dist_transform,0.7*dist_transform.max(),255,0)

# Finding unknown region
sure_fg = np.uint8(sure_fg)
unknown = cv2.subtract(sure_bg,sure_fg)

# Marker labelling
ret, markers = cv2.connectedComponents(sure_fg)

# Add one to all labels so that sure background is not 0, but 1
markers = markers+1

# Now, mark the region of unknown with zero
markers[unknown==255] = 0
markers = cv2.watershed(img,markers)
img[markers == -1] = [255,0,0]

plt.imshow(img)
plt.show()

opencv(28)---GrabCut & FloodFill图像分割 虽然是转载的，感觉还可哟