MBA interpolates irregularly and regularly spaced data using multilevel B-spline approximation. The package provides functions for estimating smooth surfaces on a regular grid and for predicting surface values at arbitrary point locations.

Installation

Install the CRAN release with:

install.packages("MBA")

Install the development version from GitHub with:

remotes::install_github("finleya/MBA")

Basic Use

library(MBA)
data(LIDAR)
set.seed(1)
train <- sample(seq_len(nrow(LIDAR)), 500)
xyz <- LIDAR[train, ]
surf <- mba.surf(xyz, no.X = 100, no.Y = 100, extend = TRUE)
image(surf$xyz.est, xaxs = "r", yaxs = "r")
pts <- mba.points(
  xyz = xyz,
  xy.est = LIDAR[-train, c("x", "y")],
  h = 8,
  verbose = FALSE
)
head(pts$xyz.est)

Functionality

The package provides:

• Surface interpolation from bivariate scattered data with mba.surf().
• Point prediction from fitted multilevel B-spline approximations with mba.points().
• Optional convex-hull masking for grid surfaces.
• Optional sp output for gridded surfaces.
• The LIDAR example dataset for interpolation examples.

Citation

The MBA implementation calls portions of the SINTEF Multilevel B-spline Library written by Oyvind Hjelle and implements methods developed in:

Lee, S., Wolberg, G., and Shin, S. Y. (1997). Scattered data interpolation with multilevel B-splines. IEEE Transactions on Visualization and Computer Graphics, 3(3), 229-244. doi:10.1109/2945.620490.