jorgemarpa/image_raw

Image Tools

This repo has a few implementations of image alignment, pattern recognition algorithms, and example of how to work with RAW images It uses OpenCV and Scikit-Image packages to open and analyze images as matrices. It uses rawpy to open RAW images and work as matrices. It uses ffmpeg-python to make time-lapse videso from a set of images.

• Houghs Transform finds edges in the image via gradient computation to then fit circles or lines. Is good to find lines and cirles in a image. I used to find the Sun disk in eclipse images and cut square stamps to center the sun and make a time-lapse. See tutorial.
• Feature Alignment first find features in the image to then find the transformation metween then with RANSAC algorthm.
• Enhanced Correlation Coeafficeint Maximization finds the transofrmation by maximizing the correlation coeafficient between the images.
• Rotation Alignment fing the best agle that minimizes resiuals in an image and applies the transformation.
• FFT phase correlation finds the best tranlation in the FFT space.

Examples

Hougs Transform to find the Sun Disk and make stamps from my 2023 Annular solar eclipse images.