Advanced gridding solutions for the petroleum, mining, GIS and oceanographic industries. Java-based Minimum Curvature Gridding and Cyclical B-Spline Gridding.

Home | Software | Publications | Company | Customers

Discussion and real time example of using Cyclical B-Spline Gridding

Cyclical B-Spline Gridding Algorithm

Cyclical B-Spline Gridding builds grid-based surface models limited only by the amount of disk space available to your computer or computer network.

The B-Spline algorithm is hierarchical, starting with a grid with 2 rows and 2 columns and though repeated refinements produces a grid at the user specified row and column spacing.

All input data is sorted on it's x coordinate and stored on disk. If not all the data can fit in memory, a merge sort algorithm is used.

During each algorithm iteration the data is read one record at a time. The equivalent of only 20 grid columns are in memory at any one time, with the rest of the grid stored on disk.

The point of keeping data and most of the grid on disk is to guarantee that grids of any desired size can be created from any amount of data, limited only by available disk space. There is reading and writing penalty paid for this extreme use of disk space for data sets and surface models that could fit in memory.

This data access penatly is worthwhile given the extremely large amounts of data available in Lidar surveys, bathymentric modeling, and basin wide modeling from seismic data.

Cyclical B-Spline Gridding does not honor faults. It's strength's are an output model that honors the data and the size of data sets it can model. In any case, faults are not needed for terrain and bathymetric modeling. And relatively fast looks at seismic data are worthwhiile even without using faults. If you data has fault discontinuties, use ASM's Java Minimum Curvature Gridding algorithm.

The images below show results of CBSG applies to Lidar point cloud data.

At the right is a video of the execution of the CBSG on this data.The example was done on a PC runnning Windows 7 which has 3 giga-bytes of memory and a dual core AMD CPU.

Lidar Point Cloud	Terrain Model Created by Running the Algorithm Described Above

Press here to read an article on CBSG.

Real Time Example Running Cyclical B-Spline Gridding

Press Play to model a lidar cloud with 2.7 million returns, and produce a grid with 28 Million Nodes

Running time is 31 Seconds on a computer using a dual core AMD CPU

A corner of the lidar data and the resulting surface model are shown at the left below

Home