Scalable Terrain Model From Point Cloud Data in Decartes

Question

Hello, 
 I am trying to create a scalable terrain model from point cloud data using Decartes in Microstation. I am attaching a pts file to the point cloud menu in microstation and converting to a pod file. The point cloud is made of 3 registered scan and the data is relatively dense. 
 The point cloud data i am importing are rock wall faces. I am setting and ACS to the wall before creating the model. 
 
 Once this is done, i create a scalable terrain model using the point cloud found in the working dgn file. Once generated, the scalable terrain model has hundred of spikes. The model has no smooth continuous areas. 
 
 Is there something i am missing on the point cloud import? Am i missing a major step in creating the scalable terrain model? Does anyone know how to create a smooth model? 
 
 Thank You!!

MichaelGilham · Accepted Answer

Have you investigated the data to see if, where there are spikes in the triangles, there are point cloud points? 
 That's' step #1. 
 Whenever a triangulation engine is attempting to triangulate data that on vertical or near vertical surfaces, there is the chance of this type of result. So take a close look at the point data and the resulting triangles and see if they match. If they do, then its a data related issue.

Mathieu St-pierre · Answer

Hi, 
 A scalable terrain model, as the name implied, is only for 2.5D data (i.e. : dataset for which each point has a unique x and y coordinates). 
 As far as I understand what you are trying to model is more 3D in nature. When triangulating 3D data points which have similar x and y coordinates will be filtered so that only 1 of those points are kept for triangulation purpose, which could lead to spike. 
 So I think you should use instead the MicroStation meshing tool, with the drawback that the number of points that can be handle in a mesh is limited. 
 Thanks, 
 Mathieu