plot.im

From spatstat v1.64-1 by Adrian Baddeley

Plot a Pixel Image

Plot a pixel image.

Keywords: hplot, spatial

Usage

# S3 method for im
plot(x, …,
                   main,
                   add=FALSE, clipwin=NULL,
                   col=NULL, valuesAreColours=NULL, log=FALSE,
                   ncolours=256, gamma=1, 
                   ribbon=show.all, show.all=!add,
                   ribside=c("right", "left", "bottom", "top"),
                   ribsep=0.15, ribwid=0.05, ribn=1024,
                   ribscale=1, ribargs=list(), riblab=NULL, colargs=list(),
                   useRaster=NULL, workaround=FALSE, zap=1,
                   do.plot=TRUE) 
  # S3 method for im
image(x, …,
                   main,
                   add=FALSE, clipwin=NULL,
                   col=NULL, valuesAreColours=NULL, log=FALSE, 
                   ncolours=256, gamma=1, 
                   ribbon=show.all, show.all=!add,
                   ribside=c("right", "left", "bottom", "top"),
                   ribsep=0.15, ribwid=0.05, ribn=1024,
                   ribscale=1, ribargs=list(), riblab=NULL, colargs=list(),
                   useRaster=NULL, workaround=FALSE, zap=1,
                   do.plot=TRUE)

Arguments

x: The pixel image to be plotted. An object of class "im" (see im.object).
…: Extra arguments passed to image.default to control the plot. See Details.
main: Main title for the plot.
add: Logical value indicating whether to superimpose the image on the existing plot (add=TRUE) or to initialise a new plot (add=FALSE, the default).
clipwin: Optional. A window (object of class "owin"). Only this subset of the image will be displayed.
col: Colours for displaying the pixel values. Either a character vector of colour values, an object of class colourmap, or a function as described under Details.
valuesAreColours: Logical value. If TRUE, the pixel values of x are to be interpreted as colour values.
log: Logical value. If TRUE, the colour map will be evenly-spaced on a logarithmic scale.
ncolours: Integer. The default number of colours in the colour map for a real-valued image.
gamma: Exponent for the gamma correction of the colours. A single positive number.
ribbon: Logical flag indicating whether to display a ribbon showing the colour map. Default is TRUE for new plots and FALSE for added plots.
show.all: Logical value indicating whether to display all plot elements including the main title and colour ribbon. Default is TRUE for new plots and FALSE for added plots.
ribside: Character string indicating where to display the ribbon relative to the main image.
ribsep: Factor controlling the space between the ribbon and the image.
ribwid: Factor controlling the width of the ribbon.
ribn: Number of different values to display in the ribbon.
ribscale: Rescaling factor for tick marks. The values on the numerical scale printed beside the ribbon will be multiplied by this rescaling factor.
ribargs: List of additional arguments passed to image.default, axis and axisTicks to control the display of the ribbon and its scale axis. These may override the … arguments.
riblab: Text to be plotted in the margin near the ribbon. A character string or expression to be interpreted as text, or a list of arguments to be passed to mtext.
colargs: List of additional arguments passed to col if it is a function.
useRaster: Logical value, passed to image.default. Images are plotted using a bitmap raster if useRaster=TRUE or by drawing polygons if useRaster=FALSE. Bitmap raster display tends to produce better results, but is not supported on all graphics devices. The default is to use bitmap raster display if it is supported.
workaround: Logical value, specifying whether to use a workaround to avoid a bug which occurs with some device drivers in R, in which the image has the wrong spatial orientation. See the section on Image is Displayed in Wrong Spatial Orientation below.
zap: Noise threshold factor. A numerical value greater than or equal to 1. If the range of pixel values is less than zap * .Machine$double.eps, the image will be treated as constant. This avoids displaying images which should be constant but contain small numerical errors.
do.plot: Logical value indicating whether to actually plot the image and colour ribbon. Setting do.plot=FALSE will simply return the colour map and the bounding box that were chosen for the plot.

Details

This is the plot method for the class "im". [It is also the image method for "im".]

The pixel image x is displayed on the current plot device, using equal scales on the x and y axes.

If ribbon=TRUE, a legend will be plotted. The legend consists of a colour ribbon and an axis with tick-marks, showing the correspondence between the pixel values and the colour map.

Arguments ribside, ribsep, ribwid control the placement of the colour ribbon. By default, the ribbon is placed at the right of the main image. This can be changed using the argument ribside. The width of the ribbon is ribwid times the size of the pixel image, where `size' means the larger of the width and the height. The distance separating the ribbon and the image is ribsep times the size of the pixel image.

The ribbon contains the colours representing ribn different numerical values, evenly spaced between the minimum and maximum pixel values in the image x, rendered according to the chosen colour map.

The argument ribargs controls the annotation of the colour ribbon. It is a list of arguments to be passed to image.default, axis and axisTicks. To plot the colour ribbon without the axis and tick-marks, use ribargs=list(axes=FALSE). To ensure that the numerals or symbols printed next to the colour map are oriented horizontally, use ribargs=list(las=1). To double the size of the numerals or symbols, use ribargs=list(cex.axis=2). To control the number of tick-marks, use ribargs=list(nint=N) where N is the desired number of intervals (so there will be N+1 tickmarks, subject to the vagaries of R internal code).

The argument riblab contains text that will be displayed in the margin next to the ribbon.

The argument ribscale is used to rescale the numerical values printed next to the colour map, for convenience. For example if the pixel values in x range between 1000 and 4000, it would be sensible to use ribscale=1/1000 so that the colour map tickmarks would be labelled 1 to 4.

Normally the pixel values are displayed using the colours given in the argument col. This may be either

an explicit colour map (an object of class "colourmap", created by the command colourmap). This is the best way to ensure that when we plot different images, the colour maps are consistent.
a character vector or integer vector that specifies a set of colours. The colour mapping will be stretched to match the range of pixel values in the image x. The mapping of pixel values to colours is determined as follows.

logical-valued images:
the values FALSE and TRUE are mapped to the colours col[1] and col[2] respectively. The vector col should have length 2.

factor-valued images:
the factor levels levels(x) are mapped to the entries of col in order. The vector col should have the same length as levels(x).

numeric-valued images:
By default, the range of pixel values in x is divided into n = length(col) equal subintervals, which are mapped to the colours in col. (If col was not specified, it defaults to a vector of 255 colours.)
Alternatively if the argument zlim is given, it should be a vector of length 2 specifying an interval of real numbers. This interval will be used instead of the range of pixel values. The interval from zlim[1] to zlim[2] will be mapped to the colours in col. This facility enables the user to plot several images using a consistent colour map.
Alternatively if the argument breaks is given, then this specifies the endpoints of the subintervals that are mapped to each colour. This is incompatible with zlim.
The arguments col and zlim or breaks are then passed to the function image.default. For examples of the use of these arguments, see image.default.
a function in the R language with an argument named range or inputs.
If col is a function with an argument named range, and if the pixel values of x are numeric values, then the colour values will be determined by evaluating col(range=range(x)). The result of this evaluation should be a character vector containing colour values, or a "colourmap" object. Examples of such functions are beachcolours and beachcolourmap.
If col is a function with an argument named inputs, and if the pixel values of x are discrete values (integer, logical, factor or character), then the colour values will be determined by evaluating col(inputs=p) where p is the set of possible pixel values. The result should be a character vector containing colour values, or a "colourmap" object.
a function in the R language with first argument named n. The colour values will be determined by evaluating col(n) where n is the number of distinct pixel values, up to a maximum of 128. The result of this evaluation should be a character vector containing color values. Examples of such functions are heat.colors, terrain.colors, topo.colors and cm.colors.

If spatstat.options("monochrome") has been set to TRUE then all colours will be converted to grey scale values.

Other graphical parameters controlling the display of both the pixel image and the ribbon can be passed through the ... arguments to the function image.default. A parameter is handled only if it is one of the following:

a formal argument of image.default that is operative when add=TRUE.
one of the parameters "main", "asp", "sub", "axes", "xlab", "ylab" described in plot.default.
one of the parameters "ann", "cex", "font", "cex.axis", "cex.lab", "cex.main", "cex.sub", "col.axis", "col.lab", "col.main", "col.sub", "font.axis", "font.lab", "font.main", "font.sub" described in par.
the argument box, a logical value specifying whether a box should be drawn.

Images are plotted using a bitmap raster if useRaster=TRUE or by drawing polygons if useRaster=FALSE. Bitmap raster display (performed by rasterImage) tends to produce better results, but is not supported on all graphics devices. The default is to use bitmap raster display if it is supported according to dev.capabilities.

Alternatively, the pixel values could be directly interpretable as colour values in R. That is, the pixel values could be character strings that represent colours, or values of a factor whose levels are character strings representing colours.

If valuesAreColours=TRUE, then the pixel values will be interpreted as colour values and displayed using these colours.
If valuesAreColours=FALSE, then the pixel values will not be interpreted as colour values, even if they could be.
If valuesAreColours=NULL, the algorithm will guess what it should do. If the argument col is given, the pixel values will not be interpreted as colour values. Otherwise, if all the pixel values are strings that represent colours, then they will be interpreted and displayed as colours.

If pixel values are interpreted as colours, the arguments col and ribbon will be ignored, and a ribbon will not be plotted.

Value

The colour map used. An object of class "colourmap".

Also has an attribute "bbox" giving a bounding box for the plot (containing the main colour image and the colour ribbon if plotted). If a ribbon was plotted, there is also an attribute "bbox.legend" giving a bounding box for the ribbon image. Text annotation occurs outside these bounding boxes.

Complex-valued images

If the pixel values in x are complex numbers, they will be converted into four images containing the real and imaginary parts and the modulus and argument, and plotted side-by-side using plot.imlist.

Monochrome colours

If spatstat.options("monochrome") has been set to TRUE, then the image will be plotted in greyscale. The colours are converted to grey scale values using to.grey. The choice of colour map still has an effect, since it determines the final grey scale values.

Monochrome display can also be achieved by setting the graphics device parameter colormodel="grey" when starting a new graphics device, or in a call to ps.options or pdf.options.

Image Looks Like Noise

An image plot which looks like digital noise can be produced when the pixel values are almost exactly equal but include a tiny amount of numerical error. To check this, look at the numerals plotted next to the colour ribbon, or compute diff(range(x)), to determine whether the range of pixel values is almost zero. The behaviour can be suppressed by picking a larger value of the argument zap.

Image Rendering Errors and Problems

The help for image.default and rasterImage explains that errors may occur, or images may be rendered incorrectly, on some devices, depending on the availability of colours and other device-specific constraints.

If the image is not displayed at all, try setting useRaster=FALSE in the call to plot.im. If the ribbon colours are not displayed, set ribargs=list(useRaster=FALSE).

Errors may occur on some graphics devices if the image is very large. If this happens, try setting useRaster=FALSE in the call to plot.im.

The error message useRaster=TRUE can only be used with a regular grid means that the $x$ and $y$ coordinates of the pixels in the image are not perfectly equally spaced, due to numerical rounding. This occurs with some images created by earlier versions of spatstat. To repair the coordinates in an image X, type X <- as.im(X).

Image is Displayed in Wrong Spatial Orientation

If the image is displayed in the wrong spatial orientation, and you created the image data directly, please check that you understand the spatstat convention for the spatial orientation of pixel images. The row index of the matrix of pixel values corresponds to the increasing $y$ coordinate; the column index of the matrix corresponds to the increasing $x$ coordinate (Baddeley, Rubak and Turner, 2015, section 3.6.3, pages 66--67).

Images can be displayed in the wrong spatial orientation on some devices, due to a bug in the device driver. This occurs only when the plot coordinates are reversed, that is, when the plot was initialised with coordinate limits xlim, ylim such that xlim[1] > xlim[2] or ylim[1] > ylim[2] or both. This bug is reported to occur only when useRaster=TRUE. To fix this, try setting workaround=TRUE, or if that is unsuccessful, useRaster=FALSE.

References

Baddeley, A., Rubak, E. and Turner, R. (2015) Spatial Point Patterns: Methodology and Applications with R. Chapman and Hall/CRC Press.

Examples

# NOT RUN {
   # an image
   Z <- setcov(owin())
   plot(Z)
   plot(Z, ribside="bottom")
   # stretchable colour map
   plot(Z, col=rainbow)
   plot(Z, col=terrain.colors(128), axes=FALSE)
   # fixed colour map
   tc <- colourmap(rainbow(128), breaks=seq(-1,2,length=129))
   plot(Z, col=tc)
   # colour map function, with argument 'range'
   plot(Z, col=beachcolours, colargs=list(sealevel=0.5))
   # tweaking the plot
   plot(Z, main="La vie en bleu", col.main="blue", cex.main=1.5,
        box=FALSE,
        ribargs=list(col.axis="blue", col.ticks="blue", cex.axis=0.75))
   # add axes and axis labels
   plot(Z, axes=TRUE, ann=TRUE, xlab="Easting", ylab="Northing")
   # log scale
   V <- eval.im(exp(exp(Z+2))/1e4)
   plot(V, log=TRUE, main="Log scale")
   # it's complex
   Y <- exp(Z + V * 1i)
   plot(Y)
# }

Documentation reproduced from package spatstat, version 1.64-1, License: GPL (>= 2)

Community examples

Looks like there are no examples yet.