拓扑和细化

感觉和测度空间有关系，暂时记下。

An包含于An-1...包含于A1.An是最终的细化骨架。

def zhangSuen(image):
    "the Zhang-Suen Thinning Algorithm"
    Image_Thinned = image.copy()  # deepcopy to protect the original image
    changing1 = changing2 = 1        #  the points to be removed (set as 0)
    while changing1 or changing2:   #  iterates until no further changes occur in the image
        # Step 1
        changing1 = []
        rows, columns = Image_Thinned.shape               # x for rows, y for columns
        for x in range(1, rows - 1):                     # No. of  rows
            for y in range(1, columns - 1):            # No. of columns
                P2,P3,P4,P5,P6,P7,P8,P9 = n = neighbours(x, y, Image_Thinned)
                if (Image_Thinned[x][y] == 1     and    # Condition 0: Point P1 in the object regions 
                    2 <= sum(n) <= 6   and    # Condition 1: 2<= N(P1) <= 6
                    transitions(n) == 1 and    # Condition 2: S(P1)=1  
                    P2 * P4 * P6 == 0  and    # Condition 3   
                    P4 * P6 * P8 == 0):         # Condition 4
                    changing1.append((x,y))
        for x, y in changing1: 
            Image_Thinned[x][y] = 0
        # Step 2
        changing2 = []
        for x in range(1, rows - 1):
            for y in range(1, columns - 1):
                P2,P3,P4,P5,P6,P7,P8,P9 = n = neighbours(x, y, Image_Thinned)
                if (Image_Thinned[x][y] == 1   and        # Condition 0
                    2 <= sum(n) <= 6  and       # Condition 1
                    transitions(n) == 1 and      # Condition 2
                    P2 * P4 * P8 == 0 and       # Condition 3
                    P2 * P6 * P8 == 0):            # Condition 4
                    changing2.append((x,y))    
        for x, y in changing2: 
            Image_Thinned[x][y] = 0
    return Image_Thinned

#转载：Skeletonization-by-Zhang-Suen-Thinning-Algorithm/thinning.py at master · linbojin/Skeletonization-by-Zhang-Suen-Thinning-Algorithm · GitHub