We have a large number (126) HBC images in our incoming collection. Of these TB has processed eleven, and marked up five so far. Those have been added to the maps folders, both locally on my machine (and therefore backed-up to Rutabaga) and on the coldesp account on home1t. This is documentation of how such images should be processed prior to being marked up with the Image Markup Tool.
- The first stage in the process is for someone to examine each image, trim it (to remove the colour bar etc.), and possibly try to enhance it, especially if it's a bit dark or murky.
- Next, resizing and creating the various versions. We constrain our full-size images to a maximum width of 5000 pixels; we also use images at 1000, 200, and 100 pixels (the last are our thumbnails). There is a script in trunk/utilities called
process_map_images.shwhich you run as follows:
- Copy all the trimmed/enhanced images you want to process into an empty folder.
- Copy the script there too, and make it executable. (The script is for Bash, and requires ImageMagick.)
- Run the script. It should create subfolders for each of the four sizes we require.
- Copy the contents of those folders into the equivalent folders where the existing maps are stored (including on home1t/coldesp/www/maps).
- An encoder can now get one of these images off the web server, and store it directly in the
trunkfolder in their svn repo (without adding it to svn, of course). Then they can clone an existing image markup file, tweak the metadata in Oxygen, then in the Image Markup Tool, delete all the boxes and replace the image. Then they're ready to start annotating.