Classification of 3D Pointclouds

[gost]

i have a Problem: i have to write my Bachelor Thesis, und the subject is "Classification of a 3D PointCloud"

Can you help me how to start?

Thanks

[daniel]

Something like:

"Since the beginning of time, man has tried to classify point clouds."

Sorry, just kidding. You'll probably have to find some recent scientific articles with an exhaustive introduction about preliminary work (you can use Google Scholar). Or if you can get a hand on a book or a PhD manuscript (look for Belton's work for instance) it would be even better.

It will be hard to find a unique source that will speak about classification of ANY point cloud. They are generally focused on a particular sensor or on a particular application.

[Fincer]

You're quite in the same situation as I am right now. However, my subject differs a little.

Anyway, have you checked these
- LASTools
- Terrasolid (-> Click "Basic Training for LiDAR Data and Images")
- CANUPO Classification - User Guide

for example?

What comes to the theory part: You may need to check the library/database of your school for scientific articles about point cloud classification. You may try to look some e-books related to the subject or a library of computer graphics to get more scientific/theoretical information.

Some useful words related especially to airborne lidar data classification are

1: Low points
2: Ground points
3: Low, medium, high vegetation points
4: Buildings
5: Model keypoints
6: Vectorized Buildings

(For example, Terrascan software uses this classification method for airborne lidar data)