Image Processing

 

When we observed the MIR station we were still learning a lot about satellite tracking and the number of usable frames from a pass was quite low - this meant than manual processing using tools like Astrostack and Paintshop Pro was not too time consuming (see full description).

However, now that we can routinely keep the ISS in frame for an entire pass we found that we needed to automate the process. We've now developed programs which perform the following steps automatically:-

bulletlocate the ISS image on a captured video frame (1500x1125) and auto crop that section to 300x300
bulletFlip the cropped image to give a "naked eye" orientation
bulletRemove background noise
bulletContrast stretch the image
bulletInsert range and time info by using a lookup table

This means that the only time consuming step is capturing the best images from the video.

October 2002 - the latest code which Phil has developed now automates the process further and can work with an entire video of a pass and produces a finished AVI animation in less that an hour. As well as the steps outlined above the software can stack a specified number of frames (typically 25 i.e. 1 second of video) and perform deconvolution to improve image clarity.

 


Previous Home Up Next

If you have any comments or suggestions about these pages please contact the author.
Page content 2001-2007 Mike Tyrrell