Improving the quality of online live video streams, super resolution methods for stars

Ben,
I am still trying to track all online live videos, particularly those that can be used for education.
This one, at the bottom of the description has a link to a Google app to set the clock.  I want to have time when you slide the slider, not hours from “now” which changes constantly.
I need a way to record or to take the frames as they are streaming.  I can do it with Visual Basic and C and MatLab. But there must be a way to read the frames from a live stream using Javascript in the browser.  I have seen some things, but I need help understanding them.  It is not fair I have to do this all by myself.
Here is another one, and another
I want to read these frames and stack the frames to get statistical average images over time.  It means shifting the pictures by tiny fragments of a pixel in each frame, getting a much higher resolution image from the many.
All that jumping around smooths out.  Some is atmosphere, some is camera and a tiny tiny bit is from the stars themselves.  I am patient.  If I observe the same set of stars night after night all night long.  It is 70 degrees across and 50 degrees vertical.  At 15 degrees rotation rate, that is three and maybe a bit more hour.  Three hours at 30 frames per second is 3600*30*3 = 324,000 pictures of each star.  I want to see what those pictures look like.
I know that the light is strong and being scattered. The atmosphere acts like a diverging statistical lens to spread the signal out so you can see it. Tracking all the pixels nearby, the halo contains information about the star and other stars nearby.  I think I can tease out much of that. And certainly improve the viewing and quality of the videos — anywhere things are not moving too much.
This is stars. But the same thing applies to buildings, mountains, moon, sun, trees, water – with variations.
I want to put a div over the whole video image and let the user hover any star to get its name, location and properties.  Later put all the images of one star seem by different videos together.  Same with moon, mars, planets, comets, sun, clouds, buildings,  things.
It is NOT hard, just tedious.  I could use some help.  Is anyone in your network interested in astronomy or machine vision, or global groups working together on YouTube videos and live videos streams?
The same methods work for stars during the day.  I want to try the same things with daytime images of the sky.  I know where the brighter stars will be, and can see how many frames and what conditions and processing are needed to know that I am seeing the stars during the day time.  Good practice for harder problems.  Some cameras will work, and others won’t.
Dad – Richard K Collins, Director, The Internet Foundation

The onscreen messages will not translate to English since they are overlays.
I always use a x10 or x20 magnifier with videos, but it is Windows magnifier and does not play well with Google Chrome and YouTube. Can you give me a magnifier window that I can put on my second screen and lock to a position in the video. For stars, can you track them and keep the magnifier centered?

I am looking at all Internet, live, all sky cameras and live telescopes, and live sensor networks for the Internet Foundation. I want to encourage ALL astronomical societies, clubs, schools, cities to have full sky cameras. I live in Houston Texas USA, and never see the stars (city lights and fog and clouds and high humidity). So I have to go online to see stars. The lossy format for YouTube videos is a problem.

I am a senior mathematical statistician. So I want to stack the frames from all sky live videos. Here is one – https://www.youtube.com/watch?v=hPWz3mDvAuY There are about 30 frames per second, and three hours when a star is on the screen. That is 324000 frames that can be averaged to improve resolution for stars, or planets, or buildings or mountains. So your timer is used for giving good time on these star videos. But if you know how to handle the video frames too, then machine vision has LOTS of applications and global communities. If you can help users like myself get access to the pixels, I can write tools to stack and display higher resolution statistically produced images and visualizations. Also, for many images an overlay of grids and coordinate systems would help. Measuring, tracking, improving resolution by stacking, combining images from different video sources. There are LOTS of traffic cameras, beach and mountain and river cameras, wildlife and place cameras. I found just over a thousand so far, but I am only getting samples. I think there are about 20,000 now. Some come and go. Many are not on YouTube. Many are unsupported and have no “support” button. If you can add twitter, you can add other buttons and sharing tools, including crowdfunding and support and thank you gifts.

Richard K Collins, Director, The Internet Foundation

I hope you will help the world share and use live videos more effectively. For education, research, scientific applications, vehicle tracking, social studies, wildlife and other things. Just ocean research, weather and astronomy (most live videos have some sky) are huge and growing. Hope you will contact me. Thanks for your app. Hope you will try some related things.

Richard K Collins

About: Richard K Collins

The Internet Foundation Internet policies, global issues, global open lossless data, global open collaboration


Leave a Reply

Your email address will not be published. Required fields are marked *