I would like to know about (zero cost) software for automatically combining a series of pixel-art images of adjacent, overlapping scenes into a single composite image.
OS isn't important. I'll be running it on Win 7 most likely, but I can use others.
These screenshots are from a game called Inherit the Earth. I've recorded the characters walking to the right up and upwards. They've been taken using the Video Capture mode in DOSBox, so I have every frame available to me as a .png sequence.
If I manually combine these together in Photoshop, I get to see more of the scene at once.
What I'd like to do is to take some thousand .pngs taken from a captured video and use software to combine all of them into a single map image of the town.
I've attempted to write a program for this in Python, but my image-comparison step seems to be the bottleneck and it takes hours to produce the following image, and some errors appear to have crept in at the center despite my efforts.
The holes are not because of mis-calibration, I hadn't walked around the entire town yet because it's big. :)
The software doesn't need to resize, rotate or flip the images, only translate.
The video smoothly scrolls, so the images are guaranteed to be adjacent to one another to within some small offset.
The images are RGB true-colour.
N.B. I'm looking for software to automatically help me do this for other games, not maps specifically for Inherit the Earth.
-- Edit - After many hours of processing and a lot of manual effort, this is the type of result I am hoping to see.