如何在python中获得voronoi边缘图

我想为点注释（原子核位置的二进制图像）创建一个voronoi边图。我使用中的代码创建voronoi边缘图，但结果显示在图像中。我只需要voronoi边。下面是代码

Voronoi边图

def viro(img):
 label_point=img
 h, w = label_point.shape
 points = np.argwhere(label_point>0)
 vor = Voronoi(points)
 regions, vertices = voronoi_finite_polygons_2d(vor)
 box = Polygon([[0, 0], [0, w], [h, w], [h, 0]])
 region_masks = np.zeros((h, w), dtype=np.int16)
 edges = np.zeros((h, w), dtype=np.bool)
 count = 1
 masks= np.zeros((h, w), dtype=np.bool)
 for region in regions:
            polygon = vertices[region]
            # Clipping polygon
            poly = Polygon(polygon)
            poly = poly.intersection(box)
            polygon = np.array([list(p) for p in poly.exterior.coords])
            mask = poly2mask(polygon[:, 0], polygon[:, 1], (h, w))            
            edge = mask * (~morphology.erosion(mask, morphology.disk(1)))
            edges += edge
            region_masks[mask] = count
            count += 1

 # fuse Voronoi edge and dilated points
 label_point_dilated = morphology.dilation(label_point, morphology.disk(1))
 label_vor = np.zeros((h, w, 3), dtype=np.uint8)
 label_vor[:, :, 0] = (edges > 0).astype(np.uint8) * 255
 label_vor[:, :, 1] = (label_point_dilated > 0).astype(np.uint8) * 255

 return label_vor    

---

我碰巧发现了这个问题。实际上，问题在于二值图像，因为它是核位置的基本事实，半径为3个像素，因此，它没有获得一个核的单坐标（单像素），而是检索了三个坐标值，因为每个核位置的半径为3个像素，从而给出了结果图像