Python cv2.getStructuringElement() Examples

def sobelOperT(self, img, blursize, morphW, morphH):
        '''
            No different with sobelOper ? 
        '''
        blur = cv2.GaussianBlur(img, (blursize, blursize), 0, 0, cv2.BORDER_DEFAULT)

        if len(blur.shape) == 3:
            gray = cv2.cvtColor(blur, cv2.COLOR_RGB2GRAY)
        else:
            gray = blur

        x = cv2.Sobel(gray, cv2.CV_16S, 1, 0, 3)
        absX = cv2.convertScaleAbs(x)
        grad = cv2.addWeighted(absX, 1, 0, 0, 0)

        _, threshold = cv2.threshold(grad, 0, 255, cv2.THRESH_OTSU + cv2.THRESH_BINARY)

        element = cv2.getStructuringElement(cv2.MORPH_RECT, (morphW, morphH))
        threshold = cv2.morphologyEx(threshold, cv2.MORPH_CLOSE, element)

        return threshold 

def findPiccircle(frame, color):

	hsv = cv2.cvtColor(frame, cv2.COLOR_BGR2HSV)	
	color_dict = color_list.getColorList()
	mask = cv2.inRange(hsv, color_dict[color][0], color_dict[color][1])
	dilated = cv2.dilate(mask, cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (3, 3)), iterations=2)
	## 需要修改minRadius以及maxRadius，用来限制识别圆的大小，排除其他的干扰
	circles = cv2.HoughCircles(dilated, cv2.HOUGH_GRADIENT, 1, 1000, param1=15, param2=10, minRadius=15, maxRadius=50)
	
	center = None
	if circles is not None:
		x, y, radius = circles[0][0]
		center = (x, y)
		cv2.circle(frame, center, radius, (0, 255, 0), 2)
		cv2.circle(frame, center, 2, (0,255,0), -1, 8, 0 );
		print('圆心：{}, {}'.format(x, y))
		
	cv2.imshow('result', frame)	
	
	if center != None:
		return center 

def maximizeContrast(imgGrayscale):

    height, width = imgGrayscale.shape

    imgTopHat = np.zeros((height, width, 1), np.uint8)
    imgBlackHat = np.zeros((height, width, 1), np.uint8)

    structuringElement = cv2.getStructuringElement(cv2.MORPH_RECT, (3, 3))

    imgTopHat = cv2.morphologyEx(imgGrayscale, cv2.MORPH_TOPHAT, structuringElement)
    imgBlackHat = cv2.morphologyEx(imgGrayscale, cv2.MORPH_BLACKHAT, structuringElement)

    imgGrayscalePlusTopHat = cv2.add(imgGrayscale, imgTopHat)
    imgGrayscalePlusTopHatMinusBlackHat = cv2.subtract(imgGrayscalePlusTopHat, imgBlackHat)

    return imgGrayscalePlusTopHatMinusBlackHat
# end function 

def skeletonize(image_in):
    '''Inputs and grayscale image and outputs a binary skeleton image'''
    size = np.size(image_in)
    skel = np.zeros(image_in.shape, np.uint8)

    ret, image_edit = cv2.threshold(image_in, 0, 255, cv2.THRESH_BINARY | cv2.THRESH_OTSU)
    element = cv2.getStructuringElement(cv2.MORPH_CROSS, (3,3))
    done = False

    while not done:
        eroded = cv2.erode(image_edit, element)
        temp = cv2.dilate(eroded, element)
        temp = cv2.subtract(image_edit, temp)
        skel = cv2.bitwise_or(skel, temp)
        image_edit = eroded.copy()

        zeros = size - cv2.countNonZero(image_edit)
        if zeros == size:
            done = True

    return skel 

def predict0():
    Vnet3d = Vnet3dModule(256, 256, 64, inference=True, model_path="model\\Vnet3dModule.pd")
    for filenumber in range(30):
        batch_xs = np.zeros(shape=(64, 256, 256))
        for index in range(64):
            imgs = cv2.imread(
                "D:\Data\PROMISE2012\Vnet3d_data\\test\image\\" + str(filenumber) + "\\" + str(index) + ".bmp", 0)
            batch_xs[index, :, :] = imgs[128:384, 128:384]

        predictvalue = Vnet3d.prediction(batch_xs)

        for index in range(64):
            result = np.zeros(shape=(512, 512), dtype=np.uint8)
            result[128:384, 128:384] = predictvalue[index]
            kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (5, 5))
            result = cv2.morphologyEx(result, cv2.MORPH_CLOSE, kernel)
            cv2.imwrite(
                "D:\Data\PROMISE2012\Vnet3d_data\\test\image\\" + str(filenumber) + "\\" + str(index) + "mask.bmp",
                result) 

def skeletize(img):
    size = np.size(img)
    skel = np.zeros(img.shape, np.uint8)
    element = cv2.getStructuringElement(cv2.MORPH_CROSS, (3, 3))
    done = False

    while not done:
        eroded = cv2.erode(img, element)
        temp = cv2.dilate(eroded, element)
        temp = cv2.subtract(img, temp)
        skel = cv2.bitwise_or(skel, temp)
        img = eroded.copy()

        zeroes = size - cv2.countNonZero(img)
        if zeroes == size:
            done = True

    return skel 

def find_components(im, max_components=16):
    """Dilate the image until there are just a few connected components.
    Returns contours for these components."""
    kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (10, 10))
    dilation = dilate(im, kernel, 6)

    count = 21
    n = 0
    sigma = 0.000

    while count > max_components:
        n += 1
        sigma += 0.005
        result = cv2.findContours(dilation, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)
        if len(result) == 3:
            _, contours, hierarchy = result
        elif len(result) == 2:
            contours, hierarchy = result
        possible = find_likely_rectangles(contours, sigma)
        count = len(possible)

    return (dilation, possible, n) 

def _get_erosion_kernel(self, mask):
        """ Get the erosion kernel.

        Parameters
        ----------
        mask: :class:`numpy.ndarray`
            The mask to be eroded or dilated

        Returns
        -------
        :class:`numpy.ndarray`
            The erosion kernel to be used for erosion/dilation
        """
        erosion_ratio = self.config["erosion"] / 100
        mask_radius = np.sqrt(np.sum(mask)) / 2
        kernel_size = max(1, int(abs(erosion_ratio * mask_radius)))
        erosion_kernel = cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (kernel_size, kernel_size))
        logger.trace("erosion_kernel shape: %s", erosion_kernel.shape)
        return erosion_kernel 

def blend_non_transparent(face_img, overlay_img):
    # Let's find a mask covering all the non-black (foreground) pixels
    # NB: We need to do this on grayscale version of the image
    gray_overlay = cv2.cvtColor(overlay_img, cv2.COLOR_BGR2GRAY)
    overlay_mask = cv2.threshold(gray_overlay, 1, 255, cv2.THRESH_BINARY)[1]

    # Let's shrink and blur it a little to make the transitions smoother...
    overlay_mask = cv2.erode(overlay_mask, cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (3, 3)))
    overlay_mask = cv2.blur(overlay_mask, (3, 3))

    # And the inverse mask, that covers all the black (background) pixels
    background_mask = 255 - overlay_mask

    # Turn the masks into three channel, so we can use them as weights
    overlay_mask = cv2.cvtColor(overlay_mask, cv2.COLOR_GRAY2BGR)
    background_mask = cv2.cvtColor(background_mask, cv2.COLOR_GRAY2BGR)

    # Create a masked out face image, and masked out overlay
    # We convert the images to floating point in range 0.0 - 1.0
    face_part = (face_img * (1 / 255.0)) * (background_mask * (1 / 255.0))
    overlay_part = (overlay_img * (1 / 255.0)) * (overlay_mask * (1 / 255.0))

    # And finally just add them together, and rescale it back to an 8bit integer image
    return np.uint8(cv2.addWeighted(face_part, 255.0, overlay_part, 255.0, 0.0)) 

def backprojection(target, roihist):
    '''图像预处理'''
    hsvt = cv2.cvtColor(target,cv2.COLOR_BGR2HSV)
    dst = cv2.calcBackProject([hsvt],[0,1],roihist,[0,180,0,256],1)
    # Now convolute with circular disc
    disc = cv2.getStructuringElement(cv2.MORPH_ELLIPSE,(7,7))
    cv2.filter2D(dst,-1,disc,dst)
    # threshold and binary AND
    ret,binary = cv2.threshold(dst,80,255,0)
    # 创建 核
    kernel = np.ones((5,5), np.uint8)
    iter_time = 1
    # 闭运算
    binary = cv2.morphologyEx(binary, cv2.MORPH_CLOSE, kernel,iterations=iter_time)

    thresh = cv2.merge((binary,binary,binary))
    target_filter = cv2.bitwise_and(target,thresh)
    
    return binary, target_filter 

def _morphological_process(image, kernel_size=5):
        """

        :param image:
        :param kernel_size:
        :return:
        """
        if image.dtype is not np.uint8:
            image = np.array(image, np.uint8)
        if len(image.shape) == 3:
            image = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)

        kernel = cv2.getStructuringElement(shape=cv2.MORPH_ELLIPSE, ksize=(kernel_size, kernel_size))

        # close operation fille hole
        closing = cv2.morphologyEx(image, cv2.MORPH_CLOSE, kernel, iterations=1)

        return closing 

def colorSearch(self, src, color, out_rect):
        """

        :param src:
        :param color:
        :param out_rect: minAreaRect
        :return: binary
        """
        color_morph_width = 10
        color_morph_height = 2

        match_gray = colorMatch(src, color, False)

        _, src_threshold = cv2.threshold(match_gray, 0, 255, cv2.THRESH_OTSU + cv2.THRESH_BINARY)

        element = cv2.getStructuringElement(cv2.MORPH_RECT, (color_morph_width, color_morph_height))
        src_threshold = cv2.morphologyEx(src_threshold, cv2.MORPH_CLOSE, element)

        out = src_threshold.copy()

        _, contours, _ = cv2.findContours(src_threshold, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_NONE)

        for cnt in contours:
            mr = cv2.minAreaRect(cnt)
            if self.verifySizes(mr):
                out_rect.append(mr)

        return out 

def spline_transform_multi(img, mask):
    bimask=mask>0
    M,N=np.where(bimask)
    w=np.ptp(N)+1
    h=np.ptp(M)+1
    kernel=cv2.getStructuringElement(cv2.MORPH_ELLIPSE,(3,3))
    bound=cv2.dilate(bimask.astype('uint8'),kernel)-bimask
    y,x=np.where(bound>0)

    if x.size>4:
        newxy=thin_plate_transform(x,y,w,h,mask.shape[:2],num_points=5)

        new_img=cv2.remap(img,newxy,None,cv2.INTER_LINEAR)
        new_msk=cv2.remap(mask,newxy,None,cv2.INTER_NEAREST)
    elif x.size>0:
        new_img=img
        new_msk=mask
    return new_img,new_msk 

def __init__(self, min_accuracy, min_blend_area, kernel_fill=20, dist_threshold=15000, history=400):
        self.min_accuracy = max (min_accuracy, 0.7)
        self.min_blend_area = min_blend_area
        self.kernel_clean = cv2.getStructuringElement(cv2.MORPH_ELLIPSE,(4,4))
        self.kernel_fill = np.ones((kernel_fill,kernel_fill),np.uint8)
        self.dist_threshold = dist_threshold
        self.history = history
   


        # read https://docs.opencv.org/3.3.0/d2/d55/group__bgsegm.html#gae561c9701970d0e6b35ec12bae149814

        try:
      
            self.fgbg = cv2.bgsegm.createBackgroundSubtractorMOG(history=self.history, nmixtures=5, backgroundRatio=0.7, noiseSigma=0)
        except AttributeError as error:
            print ('It looks like your OpenCV version does not include bgsegm. Switching to createBackgroundSubtractorMOG2')
            self.fgbg = cv2.createBackgroundSubtractorMOG2(detectShadows=False, history=self.history) 
           

        #self.fgbg = cv2.bgsegm.createBackgroundSubtractorGMG(decisionThreshold=0.98, initializationFrames=10)
        #self.fgbg = cv2.createBackgroundSubtractorMOG2(detectShadows=False, history=self.history) 
        #self.fgbg=cv2.bgsegm.createBackgroundSubtractorGSOC(noiseRemovalThresholdFacBG=0.01, noiseRemovalThresholdFacFG=0.0001)
        #self.fgbg=cv2.bgsegm.createBackgroundSubtractorCNT(minPixelStability = 5, useHistory = True, maxPixelStability = 5 *60,isParallel = True)
        #self.fgbg=cv2.createBackgroundSubtractorKNN(detectShadows=False, history=self.history, dist2Threshold = self.dist_threshold)
        #fgbg=cv2.bgsegm.createBackgroundSubtractorLSBP()

        utils.success_print('Background subtraction initialized') 

def predict0():
    Vnet3d = Vnet3dModule(256, 256, 64, inference=True, model_path="model\\Vnet3dModule.pd")
    for filenumber in range(30):
        batch_xs = np.zeros(shape=(64, 256, 256))
        for index in range(64):
            imgs = cv2.imread(
                "D:\Data\PROMISE2012\Vnet3d_data\\test\image\\" + str(filenumber) + "\\" + str(index) + ".bmp", 0)
            batch_xs[index, :, :] = imgs[128:384, 128:384]

        predictvalue = Vnet3d.prediction(batch_xs)

        for index in range(64):
            result = np.zeros(shape=(512, 512), dtype=np.uint8)
            result[128:384, 128:384] = predictvalue[index]
            kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (5, 5))
            result = cv2.morphologyEx(result, cv2.MORPH_CLOSE, kernel)
            cv2.imwrite(
                "D:\Data\PROMISE2012\Vnet3d_data\\test\image\\" + str(filenumber) + "\\" + str(index) + "mask.bmp",
                result) 

def generate_edge(label, edge_width=10, area_thrs=200):
    label_list = [7, 8, 11, 12, 13, 17, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 31, 32, 33]
    edge = np.zeros_like(label, dtype=np.uint8)
    for i in np.unique(label):
        # have no instance
        if i < 1000 or (i // 1000) not in label_list:
            continue
        
        # filter out small objects
        mask = (label == i).astype(np.uint8)
        if mask.sum() < area_thrs:
            continue
        
        rmin, rmax, cmin, cmax = _get_bbox(mask)
        mask_edge = _generate_edge(mask[rmin:rmax+1, cmin:cmax+1])
        edge[rmin:rmax+1, cmin:cmax+1][mask_edge > 0] = 255
    
    # dilation on edge
    kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (edge_width, edge_width))
    edge = cv2.dilate(edge, kernel)
    return edge 

def SeamlessClone_trimap(srcIm,dstIm,imMask,offX,offY):
    dstIm=dstIm.copy()
    bimsk=imMask>0

    new_msk=np.zeros(dstIm.shape[:2],dtype='uint8')
    new_msk[offY:offY+imMask.shape[0],offX:offX+imMask.shape[1]]=imMask

    dstIm[new_msk>0]=srcIm[imMask>0]

    kernel=cv2.getStructuringElement(cv2.MORPH_ELLIPSE,(3,3))
    bimsk=bimsk.astype('uint8')
    bdmsk=cv2.dilate(bimsk,kernel)-cv2.erode(bimsk,kernel)
    mask255=bdmsk>0
    mask255=(mask255*255).astype('uint8')

    offCenter=(int(offX+imMask.shape[1]/2),int(offY+imMask.shape[0]/2))

    if np.any(bdmsk>0):
        outputIm=cv2.seamlessClone(srcIm,dstIm,mask255,offCenter,cv2.MIXED_CLONE)
    else:
        outputIm=dstIm
        #when one object have very few pixels, bdmsk will be totally zero, which will cause segmentation fault.

    return outputIm,new_msk 

def blend_non_transparent(sprite, background_img):
    gray_overlay = cv2.cvtColor(background_img, cv2.COLOR_BGR2GRAY)
    overlay_mask = cv2.threshold(gray_overlay, 1, 255, cv2.THRESH_BINARY)[1]

    overlay_mask = cv2.erode(overlay_mask, cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (3, 3)))
    overlay_mask = cv2.blur(overlay_mask, (3, 3))

    background_mask = 255 - overlay_mask

    overlay_mask = cv2.cvtColor(overlay_mask, cv2.COLOR_GRAY2BGR)
    background_mask = cv2.cvtColor(background_mask, cv2.COLOR_GRAY2BGR)

    sprite_part = (sprite * (1 / 255.0)) * (background_mask * (1 / 255.0))
    overlay_part = (background_img * (1 / 255.0)) * (overlay_mask * (1 / 255.0))

    return np.uint8(cv2.addWeighted(sprite_part, 255.0, overlay_part, 255.0, 0.0)) 

def update(dummy=None):
        sz = cv2.getTrackbarPos('op/size', 'morphology')
        iters = cv2.getTrackbarPos('iters', 'morphology')
        opers = cur_mode.split('/')
        if len(opers) > 1:
            sz = sz - 10
            op = opers[sz > 0]
            sz = abs(sz)
        else:
            op = opers[0]
        sz = sz*2+1

        str_name = 'MORPH_' + cur_str_mode.upper()
        oper_name = 'MORPH_' + op.upper()
        st = cv2.getStructuringElement(getattr(cv2, str_name), (sz, sz))
        res = cv2.morphologyEx(img, getattr(cv2, oper_name), st, iterations=iters)

        draw_str(res, (10, 20), 'mode: ' + cur_mode)
        draw_str(res, (10, 40), 'operation: ' + oper_name)
        draw_str(res, (10, 60), 'structure: ' + str_name)
        draw_str(res, (10, 80), 'ksize: %d  iters: %d' % (sz, iters))
        cv2.imshow('morphology', res) 

def get_target_centers(img):

    # Hide buff line
    # img[0:70, 0:500] = (0, 0, 0)

    # Hide your name in first camera position (default)
    img[210:230, 350:440] = (0, 0, 0)
    gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
    # cv2.imwrite('1_gray_img.png', gray)

    # Find only white text
    ret, threshold1 = cv2.threshold(gray, 252, 255, cv2.THRESH_BINARY)
    # cv2.imwrite('2_threshold1_img.png', threshold1)

    # Morphological transformation
    kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (50, 5))
    closed = cv2.morphologyEx(threshold1, cv2.MORPH_CLOSE, kernel)
    # cv2.imwrite('3_morphologyEx_img.png', closed)
    closed = cv2.erode(closed, kernel, iterations=1)
    # cv2.imwrite('4_erode_img.png', closed)
    closed = cv2.dilate(closed, kernel, iterations=1)
    # cv2.imwrite('5_dilate_img.png', closed)

    (_, centers, hierarchy) = cv2.findContours(closed, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)
    return centers 

def open_operation(self,img):
        # 定义结构元素
        kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (5, 5))  # 矩形结构
        # kernel = cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (5, 5))   # 椭圆结构
        # kernel = cv2.getStructuringElement(cv2.MORPH_CROSS, (5, 5))     # 十字形结构
        opening = cv2.morphologyEx(img, cv2.MORPH_OPEN, kernel)  # 开运算
        cv2.imshow('Open operation', opening)
        return opening

    # 霍夫圆变换：检测表盘 

def detectTwoColors( img, capColors ):
    imgResult = np.zeros( img.shape, np.uint8 )
    imgB, imgG, imgR = cv2.split(img)
    for col in capColors:
        maskB = np.logical_and( imgB < col[0]+rad, imgB > col[0]-rad )
        maskG = np.logical_and( imgG < col[1]+rad, imgG > col[1]-rad )
        maskR = np.logical_and( imgR < col[2]+rad, imgR > col[2]-rad )
        mask = np.logical_and(maskR, maskG)
        mask = np.logical_and(mask, maskB)
        if col[1] > max(col[0],col[2]):
            imgResult[mask] = (0,255,0) # green
        else:
            imgResult[mask] = (0,128,255) # orange

    gray = cv2.cvtColor( imgResult, cv2.COLOR_BGR2GRAY )
    ret, binary = cv2.threshold( gray, 1, 255, cv2.THRESH_BINARY )

    kernel = cv2.getStructuringElement(cv2.MORPH_ELLIPSE,(7,7))
    binary = cv2.dilate( binary, kernel )
    
    cmpRG = imgR > imgG
    contours, hierarchy = cv2.findContours( binary, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE )
    result = []
    for cnt in contours:
        area = cv2.contourArea(cnt, oriented=True)
        if area < 0:
            cv2.drawContours(imgResult, [cnt], -1, (255,0,0), 2)
            maskTmp = np.zeros( (img.shape[0],img.shape[1]), np.uint8 )
            cv2.drawContours(maskTmp, [cnt], -1, 255, -1)
            mask = maskTmp > 0
            orange = np.count_nonzero(np.logical_and( mask, cmpRG ))
            # print abs(area), orange, orange/float(abs(area))
            frac = orange/float(abs(area))
            if abs(area) > 20 and (0.4 < frac < 0.7):
                M = cv2.moments(cnt)
                cx = int(M['m10']/M['m00'])
                cy = int(M['m01']/M['m00'])
                result.append( ((cx,cy), int(abs(area)), int(orange)) )

    return imgResult, result 

def reduce_noise_edges(im):
    structuring_element = cv2.getStructuringElement(cv2.MORPH_RECT, (1, 1))
    opening = cv2.morphologyEx(im, cv2.MORPH_OPEN, structuring_element)
    maxed_rows = rank_filter(opening, -4, size=(1, 20))
    maxed_cols = rank_filter(opening, -4, size=(20, 1))
    debordered = np.minimum(np.minimum(opening, maxed_rows), maxed_cols)
    return debordered 

def create_mask(self, img, color):
        img = cv2.cvtColor(img, cv2.COLOR_BGR2Lab)

        if color == 'green':
            threshold = [(20, 0, 128), (235, 128, 255)]
        elif color == 'white':
            threshold = [(100, 110, 110), (200, 140, 140)]

        else:
            raise Exception('Color undefined')
        
        mask = cv2.inRange(img, threshold[0], threshold[1])
        # kernel = cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (7,7))
        # mask =  cv2.morphologyEx(mask, cv2.MORPH_OPEN, kernel)
        # mask =  cv2.morphologyEx(mask, cv2.MORPH_CLOSE, kernel)

        mask = mask > 0

        # img = cv2.cvtColor(img, cv2.COLOR_YCR_CB2BGR)

        # thres_img = np.zeros_like(img, np.uint8)
        # thres_img[mask] = img[mask]

        binary_img = np.zeros((img.shape[0],img.shape[1]), np.uint8)
        binary_img[mask] = 255

        # cv2.imshow('img', binary_img)
        # cv2.waitKey(0)
        # exit(0)

        return mask 

def extract_geometry_vertices(mask, structure_size=(10, 10), approx_eps=0.01):
    """Extract polygon vertices from a boolean mask with the help of OpenCV
    utilities, as a numpy array

    Parameters
    ----------
    mask : numpy.array
        Image mask where to find polygons
    structure_size : tuple
        Size of the cv2 structuring artefact, as a tuple of horizontal and
    vertical pixels
    approx_eps : double
        Approximation coefficient, aiming at building more simplified polygons
    (this coefficient lies between 0 and 1, the larger the value is, the more
    important the approximation is)

    Returns
    -------
    numpy.array
        List of polygons contained in the mask, identified by their vertices
    """
    structure = cv2.getStructuringElement(cv2.MORPH_ELLIPSE, structure_size)
    denoised = cv2.morphologyEx(mask, cv2.MORPH_OPEN, structure)
    grown = cv2.morphologyEx(denoised, cv2.MORPH_CLOSE, structure)
    _, contours, hierarchy = cv2.findContours(
        grown, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE
    )
    polygons = [
        cv2.approxPolyDP(
            c, epsilon=approx_eps * cv2.arcLength(c, closed=True), closed=True
        )
        for c in contours
    ]
    return polygons, hierarchy 

def preprocess(img, threshold, show=False, kernel_size=(5, 5)):
    # 直方图局部均衡化
    clahe = cv2.createCLAHE(clipLimit=2, tileGridSize=(6, 6))
    img = clahe.apply(img)
    # 自适应阈值二值化
    dst = cv2.adaptiveThreshold(img, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C, cv2.THRESH_BINARY_INV, 127, threshold)
    # 闭运算开运算
    kernel = cv2.getStructuringElement(cv2.MORPH_CROSS, kernel_size)
    dst = cv2.morphologyEx(dst, cv2.MORPH_CLOSE, kernel)
    dst = cv2.morphologyEx(dst, cv2.MORPH_OPEN, kernel)

    if show:
        cv2.imshow('equlizeHist', img)
        cv2.imshow('threshold', dst)
    return dst 

def filter_mask (mask):
	# I want some pretty drastic closing
	kernel_close = cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (20, 20))
	kernel_open = cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (8, 8))
	kernel_dilate = cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (5, 5))

	# Remove noise
	opening = cv2.morphologyEx(mask, cv2.MORPH_OPEN, kernel_open)
	# Close holes within contours
	closing = cv2.morphologyEx(opening, cv2.MORPH_CLOSE, kernel_close)
	# Merge adjacent blobs
	dilation = cv2.dilate(closing, kernel_dilate, iterations = 2)

	return dilation 

def make_background_black(frame):
    """
    Makes everything apart from the main object of interest to be
    black in color.
    """
    logger.debug("Making background black...")

    # Convert from RGB to HSV
    frame = cv2.cvtColor(frame, cv2.COLOR_BGR2HSV)

    # Prepare the first mask.
    # Tuned parameters to match the skin color of the input images...
    lower_boundary = np.array([0, 40, 30], dtype="uint8")
    upper_boundary = np.array([43, 255, 254], dtype="uint8")
    skin_mask = cv2.inRange(frame, lower_boundary, upper_boundary)

    # Apply a series of erosions and dilations to the mask using an
    # elliptical kernel
    kernel = cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (3, 3))
    skin_mask = cv2.erode(skin_mask, kernel, iterations=2)
    skin_mask = cv2.dilate(skin_mask, kernel, iterations=2)

    # Prepare the second mask
    lower_boundary = np.array([170, 80, 30], dtype="uint8")
    upper_boundary = np.array([180, 255, 250], dtype="uint8")
    skin_mask2 = cv2.inRange(frame, lower_boundary, upper_boundary)

    # Combine the effect of both the masks to create the final frame.
    skin_mask = cv2.addWeighted(skin_mask, 0.5, skin_mask2, 0.5, 0.0)
    # Blur the mask to help remove noise.
    # skin_mask = cv2.medianBlur(skin_mask, 5)
    frame_skin = cv2.bitwise_and(frame, frame, mask=skin_mask)
    frame = cv2.addWeighted(frame, 1.5, frame_skin, -0.5, 0)
    frame_skin = cv2.bitwise_and(frame, frame, mask=skin_mask)

    logger.debug("Done!")
    return frame_skin 

def gen_trimap(alpha):
    k_size = random.choice(range(1, 5))
    iterations = np.random.randint(1, 20)
    kernel = cv.getStructuringElement(cv.MORPH_ELLIPSE, (k_size, k_size))
    dilated = cv.dilate(alpha, kernel, iterations)
    eroded = cv.erode(alpha, kernel, iterations)
    trimap = np.zeros(alpha.shape)
    trimap.fill(128)
    trimap[eroded >= 255] = 255
    trimap[dilated <= 0] = 0
    return trimap


# Randomly crop (image, trimap) pairs centered on pixels in the unknown regions. 

def gen_trimap(alpha):
    k_size = random.choice(range(1, 5))
    iterations = np.random.randint(1, 20)
    kernel = cv.getStructuringElement(cv.MORPH_ELLIPSE, (k_size, k_size))
    dilated = cv.dilate(alpha, kernel, iterations)
    eroded = cv.erode(alpha, kernel, iterations)
    trimap = np.zeros(alpha.shape)
    trimap.fill(128)
    trimap[eroded >= 255] = 255
    trimap[dilated <= 0] = 0
    return trimap


# Randomly crop (image, trimap) pairs centered on pixels in the unknown regions.