How to merge images into a single strip

Question

I would like to merge a few images into a single strip with the images next to each other. Using compositor didn't work out as I expected. Thinking a bit more about it, I would prefer not to use compositor.

And use python for this now I am wondering if the python part of blender could join images like that. Blender is not PIL, so I'm not sure if this is possible. The strip can be either horizontal or vertical.

* UPDATE * My alternative option, is to break open the pgn or jpg codec, read images as RGB data, and output a final RGB string basedupon multiple images. while possible i wonder if there are easier methods using Blender python API, can blender python do memory based like image manipulation, like addressing pixels directly to set their values ?

Answer 1

You can create an empty texture of size of the resulting image strip and alpha over all of your images over it in compositor and translate them with translate node into positions and render it:

With python:

• load in you image textures:

  import bpy
  
  def load_image(file_path):
      try:
          return bpy.data.images.load(filepath)
      except:
          raise NameError("Cannot load image %s" % filepath)
  
  image_names = ['first.png', 'second.jpg', 'third.bmp']
  images = []
  for name in image_names:
      real_path = bpy.abspath('//' + name)
      images.append(load_image(real_path))
  

• Get image dimension:

  images[0].size[0]    # resolution x
  images[0].size[1]    # resolution y
  

• Create empty image of desired resolution for example 1024x256:

  bg = bpy.ops.image.new(name="bg", width=1024, height=256, alpha=False)
  

• How to create nodes in compositor and link them:

  Controling compositor by python

• Render:

  render = bpy.context.scene.render
  # set this
  render.resolution_x = 1024
  render.resolution_y = 256
  # render image
  render.render()
  # use File_Output node or specify render.filepath to also save the result
  

Getting low level with manipulating image pixels:

# you get float values for each pixel like this:
# in png they are stored nicely in rgba order
image = bpy.data.images['SOME_NAME.PNG']
for i in range(0, len(image.pixels), 4):
    rgba = image.pixels[i:i+4]
    # this is also writable ofc

There are couple things to keep in mind though:

• Blender uses opengl internally so the origin is in the lower left corner of the image (with the y-axis pointing upwards).
• Some image programs might not save the transparency pixels
• The image might have alpha premultiplied, but this shouldn't be a concern. In that case the values are in form of [r * a, g * a, b * a, a]

Answer 2

Using the Import Images as Planes addon you can import several images and align them.

Select the images, set the offset value and materials for the plane.

Blender will then create an plane for each image and line them up on the x axis.

To line up the planes in a different axis there is a modified script that you can find here

The blenderartists.or thread for this script is here

Then just set up an orthographic camera to frame your images and set your render size to your needs.