Being a daytime software engineer and a sparetime photographer, I was looking for a nice DIY project for the next weeks. I finally decided to build a star tracker with a camera mount, so that I can take deep sky pictures with my DSLR.
While planning and layouting the components for this, I had to decide whether I want to manually adjust the rotation axis of the device so that it aligns with the rotation axis of the earth - or go the nerd way and let the system auto-arrange itself by attaching a suitable fixed USB camera, evaluate the video/images and write some software to move the device with stepper motors and screw drives until the correct orientation is reached.
As I do not want to reinvent the wheel, is there any star recognition framework that you've heard of? Preferrably open source, the language is of marginal importance as long as I can get the relevant data in and out. I'm thinking of something that I can feed with an image file and get back an array with the recognized objects with their x/y locations within the image. I did some research on that but could not find anything promising. What I need is an automatic identification of the brightest stars, so that I can check the current orientation against a reference image and get some vectors for the stepper motors. I hope I got the idea across.