R/ringSelect.R

In measuRing: Detection and Control of Tree-Ring Widths on Scanned Image Sections

ringSelect <- structure(function#Visual selection
###This function can include and exclude ring borders in plot outputs
###from \code{\link{ringDetect}} or \code{\link{plotSegments}}. The
###function is mapped by \code{\link{multiDetect}} for recursive
###processing of image sections.
                        
                        
                        ##details<<Columns in gray matrix are either
                        ##identified and stored by left-clicking the
                        ##mouse over the central axis of a gray image
                        ##in the plot output; pixel numbers of just
                        ##added ring borders are highlighted on the
                        ##gray raster. The raphics devices are
                        ##sequentially closed by right-clicking the
                        ##mouse. After a graphics device has been
                        ##closed, the graphics device of the following
                        ##segment is activated, and visual selection
                        ##on such a new segment can be
                        ##performed. Closing the graphics devices with
                        ##other procedures will stop the selection of
                        ##ring borders. This detection process can be
                        ##recursively developed on several image
                        ##section using \code{\link{multiDetect}}.
                        ##references<<Lara W., F. Bravo, and
                        ##S. Carlos. 2015. measuRing: An R package to
                        ##measure tree-ring widths from scanned images
                        ##Dendrochronologia, 34: 43-50.

(
    rdetect,##<<a list containing data frames of ring widths and ring
            ##borders such as that produced by
            ##\code{\link{ringDetect}}.
    any.col = TRUE ##<<logical. If FALSE only those column numbers in
                   ##gray matrix previouly identified as ring borders
                   ##can be selected.
)
{
    pixtypes <- rdetect[['ringBorders']]
    dim. <- attributes(rdetect)[['gray.dim']]
    plot. <- attributes(rdetect)[['opar']]
    options(warn = -1)
    
    f.inclu <- function(gray,pixtypes,any.col){
        segs <- length(dev.list())
        if(segs==0)
            stop('missing image segments',call.=FALSE)
        f.rown <- function(x)as.numeric(rownames(x))

        if(any.col){
            rowpix <- f.rown(pixtypes)
            col. <- 'black'}
        else{
                rowpix <- f.rown(pixtypes[pixtypes[,"borders"],])
                col. <- 'red'}           
        range. <- function(n)c(0,n)
        yrange <- range.(dim.[1])
        xrange <- range(rowpix)
        
        fexcleca <- function(rowpix,col.){
            par(plot.)
            par(mfg=c(2,1))
            xy <- identify(rowpix,rep(yrange[2]/2,length(rowpix)),
                           rowpix,tolerance=0.1,cex=0.55,col=col.,srt=90)
            return(xy)}
        
        xy <- list()
        for(i in 1:segs)
            repeat{
                dev.set(which=dev.list()[(segs+1)-i])
                if(length(dev.list())==1){
                    dev.set(which=dev.cur())}
                xy[[i]] <- fexcleca(rowpix,col.)
                print(xy[[i]])
                dev.off(which=dev.cur())
                break()}
        selected <- list()
        for(i in 1:segs)selected[[i]] <- xy[[i]]
        exclu <- unlist(selected)
        return(exclu)}
    
    return(f.inclu(gray,pixtypes,any.col))
    
### vector with column numbers in gray matrix of the
### identified ring borders.

}
,
ex=function(){
    ## Read one image in package folder:
    image1 <- system.file("P105_a.tif", package="measuRing")
    ## (not run) Initial diagnostic:
    detect1 <- ringDetect(image1,segs=2,marker=7)
    ##
    ## (not run) Choose other columns in gray matrix (see ringSelect);
    ## (not run) graphical devices from ringDetect should be active!
    ## (not run) Including columns:
    ##
    ## (uncomment and run):
    ## Toinc <- ringSelect(detect1)
    ## detect1 <- update(detect1, inclu = Toinc)
    ##
    ## (not run)  ring borders to be excluded:
    ## (uncomment and run):
    ## Toexc <- ringSelect(detect1,any.col = FALSE)
    ## detect1 <- update(detect1, exclu=Toexc)
    ## (not run) kill previous plot:
    graphics.off()
}
)