APOD Index

Methods and materials

For the image processing, the source code (generic C, uses linux/X11 for display) has been made available despite the fact it was only ever intended as a throwaway hack - see the README file for further details.

The software expects the input images to be in RAW format; that is, no header, just a stream of bytes, either 1 per pixel for greyscale (imdiff.c) or 3 per pixel for RGB (apod.c). You'll have to make sure the dimensions set in the source code agree with the image you're using. The way I prepared images was, roughly:-

1. Start with original main, before and after images, convert to greyscale, save as RAW (using e.g. irfanview).
2. Process with imdiff to create big raw images.
3. Extract relevant portions, perform rotation etc using 'gimp' or Serif Photoplus 5
4. Convert to appropriate format, again using irfanview.

src-0.0.1.tgz - 19485 bytes released 19-Dec-2004
src-0.0.2.tgz - 25959 bytes released 2-Jan-2005

For details of how the image trail and streak were replicated, see the page which reports the results - there is a description there. Also see the trail fitting results, both initial and final versions.

APOD Index