Weird triangulation output

[Ylannl]

I’m seeing some weird startin output. My code reads a LAS file with 250k points, does thinning , triangulates, and writes to OBJ.

If I set the thinning factor to ’20’ startin gives identical results compared to scipy.

However, if I set the thinning factor to ’5’ startin gives a broken TIN.

Below are the scipy and startin output for a thinning of '5'

[hugoledoux]

With thinning=5 there was a duplicate vertex, and you used the original index of points still containing the duplicate...

You need to fetch the "correct" vertices with dt.all_vertices() and then flush the infinity point (at index 0) to make the OBJ:

def write_obj_startin(pts, obj_filename):
    print("=== Writing {} ===".format(obj_filename))
    f_out = open(obj_filename, 'w')
    # for p in pts:
        # f_out.write("v {:.2f} {:.2f} {:.2f}\n".format(p[0], p[1], p[2]))
    dt = startin.DT()
    dt.insert(pts)
    pts = dt.all_vertices()
    for p in pts[1:]: 
        f_out.write("v {:.2f} {:.2f} {:.2f}\n".format(p[0], p[1], p[2]))
    for simplex in dt.all_triangles():
        f_out.write("f {} {} {}\n".format(simplex[0], simplex[1], simplex[2]))
    f_out.close()

But this is way too messy, I'll expose the OBJ export function and document this in 0.4.7 release, coming to a computer near you this week.

[Ylannl]

Ofcourse 😅🤦🏻‍♂️ That makes sense. Cheers for looking into that.