Selecting contiguous, irregularly-shaped sets of values from arrays

Dear Bryan,
 
You could try using spatial techniques to choose the contiguous areas of your matrices.
 

require(spatstat)
set.seed(1520)
x <- matrix(rnorm(25), nrow=5, ncol=5,

+          dimnames=list(c("A","B","C","D","E"), c("v","w","x","y","z")))

x

           v            w          x         y           z
A  0.3089046 -0.003350135 -0.4506777 0.7971787 -1.95078919
B -1.5895009  0.336233539 -0.3237293 0.7676754 -0.76756928
C -1.0324022 -1.119037223 -1.1525350 0.6057773 -0.28930702
D -0.8440912 -0.499994418  0.7664473 0.6367184 -0.09801227
E  0.6261038  0.391232210  0.4967601 1.0753439 -0.50998559

z <- y <- x

y[, ] <- x < 0.6
y

  v w x y z
A 1 1 1 0 1
B 1 1 1 0 1
C 1 1 1 0 1
D 1 1 0 0 1
E 0 1 1 0 1

yi <- as.im(y)
ycOut <- connected(yi, background = 0)
yc <- ycOut$v
yc

     [,1] [,2] [,3] [,4] [,5]
[1,] 1    1    1    <NA> 2  
[2,] 1    1    1    <NA> 2  
[3,] 1    1    1    <NA> 2  
[4,] 1    1    <NA> <NA> 2  
[5,] <NA> 1    1    <NA> 2  
Levels: 1 2

z[yc != 1 | is.na(yc)] <- NA
z

           v            w          x  y  z
A  0.3089046 -0.003350135 -0.4506777 NA NA
B -1.5895009  0.336233539 -0.3237293 NA NA
C -1.0324022 -1.119037223 -1.1525350 NA NA
D -0.8440912 -0.499994418         NA NA NA
E         NA  0.391232210  0.4967601 NA NA
 
Hope this helps.
 
Best wishes,
 
Chris Campbell
MANGO SOLUTIONS
Data Analysis that Delivers
+44 1249 767700
 
 
-----Original Message-----
From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-project.org] On Behalf Of Bryan McCloskey
Sent: 31 January 2012 19:58
To: r-help at r-project.org
Subject: [R] Selecting contiguous, irregularly-shaped sets of values from arrays
 
All,
 
I am attempting to select all of the contiguous elements of a matrix that meet some criterion. I.e., values that would be contained within an irregular area defined by a "contour" applied around point of interest. So, if I have a matrix x as follows:
 

x <- matrix(rnorm(25), nrow=5, ncol=5,

         dimnames=list(c("A","B","C","D","E"), c("v","w","x","y","z"))

x

           v          w          x         y          z
A  0.5184795  1.9641285  0.8632044 1.5010397  0.8468490 B -1.2402866  0.5211307 -0.1474351 1.3264893  0.1087390 C  0.5910275 -1.1708906  0.9440755 1.0970971 -0.2784806 D  0.6377495  1.1594035 -0.4217621 1.4021680 -0.6487677 E -1.4590833  0.2065765  0.1623669 1.3598283  0.3742821

 
how can I select all values in the "pond" of contiguous matrix entries that have values, say <0.6, if my entry of interest is x["A","v"]. In that case, I would like to select the following starred entries:
 
           v          w          x         y          z
A  0.5184795*  1.9641285  0.8632044 1.5010397  0.8468490 B -1.2402866*  0.5211307* -0.1474351* 1.3264893  0.1087390 C  0.5910275* -1.1708906*  0.9440755 1.0970971 -0.2784806 D  0.6377495  1.1594035 -0.4217621* 1.4021680 -0.6487677 E -1.4590833  0.2065765*  0.1623669* 1.3598283  0.3742821
 
But I would _not_ like to select any of the values in x[,"z"], because, even though they may be <0.6, they are not contiguous with the pond that x["A","v"] is in.
 
Is there an easy way to do this for many points of interest in a large matrix?
 
Thanks,
-bryan
 
------
Bryan McCloskey, Ph.D.
U.S. Geological Survey
St. Petersburg Coastal & Marine Science Center St. Petersburg
------