bp: Box-Pierce-Type Statistic

In spc4sts: Statistical Process Control for Stochastic Textured Surfaces

Box-Pierce-Type Statistic

Description

Compute a Box-Pierce-type (BP) statistic for pixels in a given image. bp2() cannot be used for pixels with the boundary problem, but is more efficient than bp() for other pixels.

Usage

bp(img, i1, i2, w, K)
bp2(img, i1, i2, w , K)

Arguments

img	the given image
i1	the row index of the pixel to compute the BP statistic for.
i2	the column index of the pixel to compute the BP statistic for.
w	the dimension of the spatial (square) moving window of the BP statistic. Must be an odd number >= 3.
K	the weighted (kernel) matrix.

Value

The BP statistic.

Warning

For pixels with the boundary problem, bp() must be used.

Note

bp() is only used in sms() for pixels with the boundary problem. It is less efficient than bp2() for other pixels.

Author(s)

Anh Bui

References

Bui, A.T. and Apley., D.W. (2018a) "A Monitoring and Diagnostic Approach for Stochastic Textured Surfaces", Technometrics, 60, 1-13.

See Also

kerMat, spaCov, sms, ad

Examples

img <- matrix(rnorm(100),10,10)
w <- 3
K <- kerMat((w + 1)/2)
## for pixels with the boundary problem, e.g., Pixel (5,1),
# running bp2(img,5,1,w,K) will produce an error; instead, use bp() in this case:
bp(img,5,1,w,K)

## for pixels without the boundary problem, e.g., Pixel (5,5),
# both can be used, but bp2() is more efficient than bp()
bp2(img,5,5,w,K)
bp(img,5,5,w,K)

spc4sts documentation built on May 24, 2022, 5:07 p.m.