Is it possible to use imagemagic to build a composite image from multiple photos of a moving object?

Question

I frequently have 10 photos taken from a tripod mounted camera, of an object moving across the field of vision.

I know I can painstakingly combine these in PhotoShop/gimp to create one composite image showing the object 10 times in a single frame.

I feel that this should be programatically possible though, by analysing the 10 photos one can find what the steady-state background is, and then overlay only the moving object from each photo on top of it?

And generally, if something's possible it means someone much cleverer than me will have already done it!

I've read all the options on the composite functionality though, and can't see anything relevant there? Google hasn't thrown up anything useful, but I may be being blind.

Many thanks

Answer 1

I'm not super familiar with ImageMagick, but have done this sort of thing in NetPpm. You would want to do something like this:

For each image in the set:

• increase pixel depth to 16 or 24 bits per channel. (48 or 72 bits/pixel)
• multiple each channel by 1/n where n is the number of images in the set.
• add the pixels together to make a new image
• reduce the image bit depth to your choice.

Increasing the bit depth will prevent banding when you do the pixel arithmetic.

This assumes that your tripod is stable enough that the images are perfectly in register. You can check this by subtracting one image from another. The only differences should be the object and the sensor noise.

This also assumes a bright object against a dark background. Ghost images of the bat-mobile racing across the Bonneville salt flats are going to be trickier.