in ,

Artificial intelligence helps NASA improve its eyes on the sun

Florida News Times

This image shows the seven UV wavelengths observed by the atmospheric imaging assembly onboard NASA’s Solar Dynamics Observatory. The upper row shows the observations from May 2010, and the lower row shows the observations from 2019 without modification, showing how the equipment deteriorated over time. Credits: Luiz Dos Santos / NASA GSFC

A group of researchers are using artificial intelligence technology to calibrate parts of NASA’s images of the sun to help scientists improve the data they use to study the sun.New technology published in journal Astronomy and astrophysics April 13, 2021.

The solar telescope does a lot of work. Staring at the sun comes at a harsh sacrifice with the endless flow of solar particles and the constant impact of strong sunlight. Over time, the sensitive lenses and sensors of the solar telescope begin to deteriorate. To ensure that the data sent back by such equipment is still accurate, scientists regularly recalibrate to make sure they understand how the equipment is changing.

Launched in 2010, NASA’s Solar Dynamics Observatory (SDO) has been providing high-resolution images of the Sun for over a decade. The image gives scientists details of various solar phenomena that can cause space weather and affect astronauts and technology on Earth and space. The Atmospheric Imagery Assembly (AIA) is one of SDO’s two imaging devices that constantly monitor the sun and take images over 10 wavelengths of UV light every 12 seconds. This creates a wealth of information about the sun like no other, but like all sun-staring instruments, the AIA deteriorates over time and requires frequent calibration of the data.

Since the launch of SDO, scientists have Sounding rocket Calibrate the AIA. Sounding rockets are small rockets that typically carry only a few devices and make short flights into space (usually only 15 minutes). What is important is that the sounding rocket flies over most of the Earth’s atmosphere so that the onboard equipment can see the UV wavelengths measured by the AIA. Light of these wavelengths is absorbed by the Earth’s atmosphere and cannot be measured from the ground. To calibrate the AIA, they attach a UV telescope to the sounding rocket Compare that data with AIA measurements. Scientists can then make adjustments to account for changes in AIA data.

The sounding rocket calibration method has some drawbacks. Sounding rockets can be launched very often, but the AIA is always looking at the sun. This means that there is a slight downtime between the sounding rocket calibrations.

Dr. Luis Dos Santos, a solar physicist and lead author of the treatise at NASA’s Goddard Space Flight Center in Greenbelt, Maryland, said: “We are working on two issues at once.”

Virtual calibration

With these challenges in mind, scientists decided to consider other options for calibrating the instrument, aiming for constant calibration. Machine learning, the technology used in artificial intelligence, seemed to fit perfectly.

As the name implies, machine learning requires computer programs or algorithms to learn how to perform tasks.

The top line of the image shows the degradation of AIA’s 304 angstrom wavelength channel over the years since the launch of SDO. The bottom line of the image uses a machine learning algorithm to compensate for this degradation. Credits: Luiz Dos Santos / NASA GSFC

First, the researchers Machine learning algorithm How to recognize the structure of the sun and compare them using AIA data. To do this, they provide algorithmic images from the sounding rocket calibration flight and convey the correct amount of calibration required. After doing these examples well, provide a similar image to the algorithm to see if you can identify the correct calibration you need. With enough data, the algorithm learns to determine the amount of calibration required for each image.

Since the AIA sees the sun in multiple wavelengths of light, researchers can also use algorithms to compare specific structures across wavelengths and enhance their assessment.

First, what will they teach the algorithm Solar flare It looked like that by displaying solar flares at all wavelengths of the AIA until we recognized solar flares in different types of light. Once the program is able to recognize solar flares without degradation, the algorithm can determine how much degradation is affecting the current image of the AIA and the amount of calibration required for each.

“This was a big deal,” said Dos Santos. “Rather than identifying it at the same wavelength, we identify the structure across wavelengths.”

This means that researchers can be more confident in calibrating the identified algorithms. In fact, the machine learning program was right when comparing their virtual calibration data with the sounding rocket calibration data.

With this new process, researchers are ready to constantly calibrate AIA images during the flight of the calibration rocket to improve the accuracy of SDO data for researchers.

Machine learning beyond the sun

Researchers are also using machine learning to better understand near-home conditions.

A group of researchers led by Dr. Ryan McGranaghan, Principal Data Scientist and Aerospace Engineer at ASTRALLC and NASA Garden Space Flight Center, used machine learning to better understand the relationship between the Earth’s magnetic field and the ionosphere, the charged part of the Earth’s upper part. I understand it well. atmosphere. Using data science technology for large amounts of data, applying machine learning technology to develop new models and how energized particles from space fall into the Earth’s atmosphere, which drives space weather. I have come to understand better.

As Machine learning As it progresses, its scientific application will expand to more and more missions. In the future, this could mean a deep space mission. calibration Rocket flight is not possible. You can calibrate and continue to provide accurate data as you move further and further away from the Earth and stars.

Sounding rocket sees the sun again

For more information:
Luiz FG Dos Santos et al, Multi-Channel Automatic Calibration of Atmospheric Imaging Assembly Using Machine Learning, Astronomy and astrophysics (2021). DOI: 10.1051 / 0004-6361 / 202040051

Quote: Artificial intelligence is NASA’s Sun acquired on July 24, 2021 from (July 24, 2021) Helps improve eyes on the day)

This document is subject to copyright. No part may be reproduced without written permission, except for fair transactions for personal investigation or research purposes. The content is provided for informational purposes only.

Source link

What do you think?


Leave a Reply

Your email address will not be published.

      Bitcoin Default Digital Reserve Currency

      Bitcoin Default Digital Reserve Currency

      Nick Wittgren reveals the death threats he’s received via social media