local_stat: Bundle of statistics, which is applied to each cell of multi-band image.

Description

If bands of ursaRaster object are interpreted as observations in time, then local_stat returns some parameters for time-series analysis. This is a local operation of map algebra.

Usage

local_stat(obj, time = NULL, cover = 1e-06, smooth = FALSE, verbose = FALSE)

Value

Object of class ursaRaster with bands:

mean: Mean value in each cell across all bands of source raster.
sd: Standard deviation in each cell across all bands of source raster. Denominator is n.
sum: Sum value in each cell across all bands of source raster.
min: Minimal value in each cell across all bands of source raster.
max: Maximal value in each cell across all bands of source raster.
n: Number of non-NA values in each cell across all bands of source raster (number of observations).
slope: Slope value in each cell across all bands of source raster.
slopeS: Significance of slope value taken with a sign of slope.
RSS: Resisual sum of squares.
ESS: Explained sum of squares.

Arguments

obj: Object of class ursaRaster
time: Numeric or NULL. If NULL then regression parameters are for regular time-series using position of band in the brick (or, time=seq(obj)). If numeric, then length of time should be equal to number of bands of obj, and time is used to set irregularity for time-series.
cover: Numeric. 0<=cover<=1 or >1. Quota for NA values in the location for all bands. Quota exceeding leads to recording NA value in the cell of created band. If code>1 then number of bands. If 0<=cover<=1 then proportion cover to number of bands. Default is 1e-6.
smooth: Logical. If TRUE then median focal smoothing is applying to created 'slope' band; it is more suitable for visualization. Default is FALSE.
verbose: Logical. Value TRUE provides some additional information on console. Default is FALSE.

Author

Nikita Platonov platonov@sevin.ru

Examples

Run this code

session_grid(NULL)
set.seed(353)
session_grid(regrid(mul=1/8))
a <- ursa_dummy(nband=15)
a[a<60] <- NA
cvr <- 12
b <- local_stat(a,cover=cvr)
print(b)
c.mean <- c(' mean'=b["mean"]
           ,'local_mean'=local_mean(a,cover=cvr)
           ,' mean'=mean(a,cover=cvr))
c.max <- c(' max'=b["max"]
          ,'local_max'=local_max(a,cover=cvr)
          ,' max'=max(a,cover=cvr))
print(c.mean)
print(c.max)
cmp <- c(mean=b["mean"]-local_mean(a,cover=cvr)
        ,sd=b["sd"]-local_sd(a,cover=cvr))
print(round(cmp,12))
d <- as.list(b)
d[["slopeS"]] <- colorize(d[["slopeS"]],stretch="signif")
display(d,blank.density=20,blank.angle=c(-45,45))