Since it is the monsoon season again, I've been thinking of writing something about astrophotography rather than actually engaging in it. Coincidentally, one of my friends sent in an article based on some scholarly work where the team was working on combining deep space photos that are general public to create more detailed photos which are comparable with the ones that are taken from the most cutting edge telescopes.
This is infact, what we all do as amateur astrophotographers. However the challenge for combining images from a large number of sources lies in the fact that you don't know what conditions they were shot under and whether all of them are of comparable quality.In this research, the team managed to set up a system where it queries for popular image search services for a given object, runs the images through a 'space locationing' software to get the accurate corrdinates, stacks them and enhances the image.
The researchers claim that due to the sheer number of images, the final image is comparable to the images produced by the best telescopes available. In my opinion, they are comparable, but not equal. However I think the research does bring in a very valid point; crowdsourcing allows to parallelise and prioritise the next target.
In one of the images, the researchers show an interesting feature known as a stellar stream on a particular galaxy (NGC 5907). The more important point is that when they processed the images for this galaxy, they handpicked and removed the images that showed any visible signs of the stellar stream. Yet the processed image that is made of a large number of images, show the stellar stream. The reason for this is the data that is actually present in the image, but not visible to the eye. By super imposing thousands of images, you can enhance the fainter details. This in it self allows early detection. Imagine this, if we were to point the best telescopes we got at all possible targets in the night sky, then it is going to cost considerable time and money. But what if we distribute the work load? Let thousands of amateur astrophotographers to do their thing and contribute if they want, process all of that information, and then direct the stronger telescopes to further inspect? I think that is the true value this research has found.
If someone is interested in the news story, you can find it at: http://www.technologyreview.com/view/528536/how-crowdsourced-astrophotographs-on-the-web-are-revolutionizing-astronomy.
The original research paper is available at the Cornell University website: http://arxiv.org/abs/1406.1528. You can download the PDF of the paper free.
This is infact, what we all do as amateur astrophotographers. However the challenge for combining images from a large number of sources lies in the fact that you don't know what conditions they were shot under and whether all of them are of comparable quality.In this research, the team managed to set up a system where it queries for popular image search services for a given object, runs the images through a 'space locationing' software to get the accurate corrdinates, stacks them and enhances the image.
The researchers claim that due to the sheer number of images, the final image is comparable to the images produced by the best telescopes available. In my opinion, they are comparable, but not equal. However I think the research does bring in a very valid point; crowdsourcing allows to parallelise and prioritise the next target.
In one of the images, the researchers show an interesting feature known as a stellar stream on a particular galaxy (NGC 5907). The more important point is that when they processed the images for this galaxy, they handpicked and removed the images that showed any visible signs of the stellar stream. Yet the processed image that is made of a large number of images, show the stellar stream. The reason for this is the data that is actually present in the image, but not visible to the eye. By super imposing thousands of images, you can enhance the fainter details. This in it self allows early detection. Imagine this, if we were to point the best telescopes we got at all possible targets in the night sky, then it is going to cost considerable time and money. But what if we distribute the work load? Let thousands of amateur astrophotographers to do their thing and contribute if they want, process all of that information, and then direct the stronger telescopes to further inspect? I think that is the true value this research has found.
If someone is interested in the news story, you can find it at: http://www.technologyreview.com/view/528536/how-crowdsourced-astrophotographs-on-the-web-are-revolutionizing-astronomy.
The original research paper is available at the Cornell University website: http://arxiv.org/abs/1406.1528. You can download the PDF of the paper free.
Comments
Post a Comment