ExifTool functionality from R

ExifTool is an excellent and comprehensive open source utility for reading, writing and editing meta information in a wide variety of files. As noted on the project homepage:

ExifTool supports many different metadata formats including EXIF, GPS, IPTC, XMP, JFIF, GeoTIFF, ICC Profile, Photoshop IRB, FlashPix, AFCP and ID3, as well as the maker notes of many digital cameras by Canon, Casio, DJI, FLIR, FujiFilm, GE, GoPro, HP, JVC/Victor, Kodak, Leaf, Minolta/Konica-Minolta, Motorola, Nikon, Nintendo, Olympus/Epson, Panasonic/Leica, Pentax/Asahi, Phase One, Reconyx, Ricoh, Samsung, Sanyo, Sigma/Foveon and Sony.

The exiftoolr R package provides a thin wrapper around a local installation of ExifTool, giving its users easy access to ExifTool functionality from within R. exiftoolr can be configured to use any ExifTool installation, but many or most users will prefer to simply run install_exiftool() once following package installation to download and install the most up-to-date version of ExifTool into the exiftoolr package directory. (install_exiftool() can also be run subsequently, at any time, to update the installed ExifTool to the most recent version.)

Once the package is installed and set up, a simple call to exif_read() can be used to read metadata from one or more files into a data.frame with one column per metadata field and one row per file. Another function, exif_call(), supports more general calls to the underlying ExifTool utility, examples of which are displayed here.

Example

First up, an example of what ExifTool (via exiftoolr) can do. Say that you have an image onto which you would like to add a bit of text indicating the time and place at which it was taken. Here is a photo taken in the LaSal mountains in southeastern Utah, USA.

Here is one way to do that, using exiftoolr to extract the relevant data, and the (excellent) magick package to annotate the image with it:

library(exiftoolr)
library(magick)

## Read and extract image metadata
dat <- exif_read("LaSals.jpg")
DateTime  <- dat[["CreateDate"]]
Longitude <- dat[["GPSLongitude"]]
Latitude  <- dat[["GPSLatitude"]]

## Prepare annotation text
txt <- paste0(DateTime, "\n",
              "Longitude: ", round(Longitude, 5), "\n",
              "Latitude:  ", round(Latitude, 5))

## Annotate image and write to file
out <- image_annotate(image_read(infile), txt,
                      gravity = "northwest", color = "red",
                      boxcolor = adjustcolor("black", alpha=0.2),
                      size = 15, location = "+10+10")
image_write(out, "LaSals_annotated.jpg")

Installation and setup

To install exiftoolr from GitHub, do:

if(!require(devtools)) {install.packages("devtools")}
devtools::install_github("JoshOBrien/exiftoolr")

Then, to install a copy of the ExifTool library or executable into the exiftoolr package's directory tree, where calls to functions in the exiftoolr package will automatically find it, run install_exiftool():

exiftoolr::install_exiftool()

exiftoolr makes a reasonable attempt to find local copies of Perl and ExifTool, and in most cases will need no hints to find them. For situations in which want more control over which Perl or ExifTool is used, set their paths with either an explicit call to configure_exiftool() or by setting the environment variables "ET_PERL_PATH" and "ET_EXIFTOOL_PATH".

Usage

exif_read() will quickly read metadata from one or more image files, returning the results in a plain data.frame:

library(exiftoolr)
image_files <- dir(system.file("images", package = "exiftoolr"), 
                   full.names = TRUE)
exifinfo <- exif_read(image_files)
#> Using ExifTool version 11.49
dim(exifinfo)
#> [1]  2 99
names(exifinfo)[1:20] ## List the first 20 metadata fields read by ExifTool
#>  [1] "SourceFile"        "ExifToolVersion"   "FileName"          "Directory"         "FileSize"         
#>  [6] "FileModifyDate"    "FileAccessDate"    "FileCreateDate"    "FilePermissions"   "FileType"         
#> [11] "FileTypeExtension" "MIMEType"          "ExifByteOrder"     "Make"              "Model"            
#> [16] "Orientation"       "XResolution"       "YResolution"       "ResolutionUnit"    "Software"

As you can see, there are a lot of columns! To extract only those tags that are actually needed, use the tags argument:

exif_read(image_files, tags = c("filename", "imagesize"))
#>                                 SourceFile   FileName ImageSize
#> 1 C:/R/Library/exiftoolr/images/LaSals.jpg LaSals.jpg   640 480
#> 2 C:/R/Library/exiftoolr/images/Lizard.jpg Lizard.jpg 4032 3024

The tags argument also accepts simple regular expressions. For instance, to extract all fields with names containing the substring "GPS", you could use the following call:

exif_read(image_files[1], tags = "*GPS*")
#>                                 SourceFile GPSLatitudeRef GPSLongitudeRef GPSAltitudeRef GPSTimeStamp GPSSpeedRef
#> 1 C:/R/Library/exiftoolr/images/LaSals.jpg              N               W              0     23:05:36           K
#>   GPSSpeed GPSImgDirectionRef GPSImgDirection GPSDestBearingRef GPSDestBearing GPSDateStamp GPSHPositioningError
#> 1        0                  T        107.2073                 T       107.2073   2016:09:21                    5
#>   GPSAltitude          GPSDateTime GPSLatitude GPSLongitude                        GPSPosition
#> 1    2257.414 2016:09:21 23:05:36Z    39.64798    -111.3705 39.6479805555556 -111.370505555556

To access more general ExifTool functionality (many examples of which are shown here), you can use the function exif_call() to roll your own ExifTool call. For the previous example, it would look something like this:

exif_call(args = c("-n", "-j", "-q", "-*GPS*"), path = image_files[1])

Why another R package for reading image file metadata?

Several existing R packages can read EXIF metadata from image files. The exif, exiv, and magick packages, all include functions (exif::read_exif(), exiv::read_exif(), and magick::image_attributes(), respectively) that extract files' EXIF data. In many files, though, EXIF tags comprise only a subset of the metadata, and even well as EXIF maker notes and (despite its name) ExifTool reads data stored in many additional metadata formats.

The exifr package (of which exiftoolr is essentially a much modified fork) is much more similar in the functionality that it provides. Both packages construct a thin wrapper around ExifTool, and so inherit its excellent support for a large variety of file types and many metadata formats beyond EXIF tags. The packages differ mainly in their support for easy installation and configuration on all operating systems. exiftoolr, in particular, was designed to make it as easy for Windows users -- even those without Python installations -- to access ExifTool functionality as it is for *NIX and Mac users. Relative to exifr, exiftoolr also makes it easier to update ExifTool to its most current version.