树的 3d 点云中的 Alphashapes

Question

I need help finding the smallest enclosing volume of the crown of a tree (forest tree) and the vertices of the points that compose that volume. This can be solved by finding an alpha shape (a concave hull) for that tree. I don't know c++ so CGAL and PCL are out of question. Is there a handy way to do this in python?

I have used ConvexHull from scipy

dataframe = pd.read_csv(base + direct + '\\' + file)
      #dataframe = pd.read_csv('C:\\Users\\jpps_\\Documents\\Projeto_Novo\\Cylinders_Eucaliptos_Pinheiros_Veg_Acima_15m\\cilnr19.0.csv')
        trees= dataframe[(dataframe['raw_classification']!=30) & (dataframe['intensity']!=64)][['X','Y','Z','intensity']].reset_index(drop=True)
        points = np.array(trees[['X','Y','Z']].values)            
        hull = ConvexHull(points)
        indices = hull.vertices
        vertices = points[indices]
        vert_df = pd.DataFrame(vertices,columns = ['X','Y','Z'])

I was hoping there is a similar procedure, but that will render me an alpha shape.

point cloud img:

Answer 1

Yes, have a look at Alpha Shape Toolbox at pypi.org :

---

^{Plate Carree Projection}

---