Module 4: How can programming and data be used to build intelligence in computers to recognize objects from space?

This module explores how programming and data can be used to build intelligence in computers that recognize objects from space. In Module 3, you learned how Satellites stay in orbit, the kinds of sensors they have, and the types of things they look at.

Module 4 – Middle School

Part 1 – What is remote sensing, and how does it work?

Part 2 – How are computers programmed to recognize objects?

Part 3 – How does artificial intelligence help computers recognize objects from space?

Part 1 – What is remote sensing, and how does it work?

Remote Sensing is the observation of something from a distance using sensors. In this case, the something is the Satellite, and the sensors it carries depend on what it has to look at. Satellites help scientists, governments, and companies collect data on the Earth. The quality and type of data the satellite collects depend on a couple of things—the orbit, the sensors, and the resolution of the sensors.

Orbits are the parking spot for the satellite over the Earth. There are three main types of orbits satellites are parked at, low-Earth orbit (about 100 miles to 1200 miles or 160 to 2,000 km above Earth), medium-Earth orbit (about 1200 miles to 22,000 miles or 2,000 to 35,500 km above Earth), and high-Earth orbit (above 22,000 miles about 35,500 km above Earth). Even though this seems a long way away, the moon is almost 239,000 miles away!! These satellites are pretty close to the Earth compared to where the moon is.

Satellites that orbit 35,786 km above the Earth are at a height(altitude) where satellite orbital speed matches the planet’s rotation, and they are in a geosynchronous orbit (GSO), meaning they are in sync with Earth’s orbit. If a geosynchronous satellite is parked over the equator it has a geostationary orbit, orbit enables a satellite to maintain its position directly over the same place on Earth’s surface. These are high Earth Orbit satellites.

A spacecraft in a geostationary orbit. Image Credit: NASA Science

The most common orbit for satellites is the low-Earth orbit since it can be programmed to follow different pathways around the planet. Polar-orbiting satellites are generally tilted 90 degrees towards the equator and travel from pole to pole as Earth rotates. (so as the Earth rotates, the satellite goes up and down the longitudinal lines) Sensors on the satellite acquire data for the whole planet, which makes them a good choice for collecting data over time. Some of the polar satellites are also Sun-Synchronous, which means that they travel over the same point on the Earth at the same time every day. If a satellite is non-polar, it does not travel the whole globe like the polar ones; it only covers a set pattern over the Earth.

Medium Orbit satellites are at about half the distance of the high Earth so they take 12 hours to complete an orbit. That means they cross the same spot on Earth twice a day. The Orbit they take is programmed, and these satellites are the most used for phone applications and GPS satellites. Starlinks are also an example of these kinds of satellites.

Energy is used to collect and transmit information

Electromagnetic energy travels as a wave through space and into our atmosphere. These waves have different energy (amplitude) and wavelengths (distance from crest to crest or trough to trough). If you have short wavelengths, that means more waves pass through a set amount of time, so you have a higher frequency. Radio waves and infrared waves

Electromagnetic energy, produced by the vibration of charged particles, travels in the form of waves through the atmosphere and the vacuum of space. These waves have different wavelengths (the distance from wave crest to wave crest) and frequencies; a shorter wavelength means a higher frequency. Radio waves and microwaves have long wavelengths and low frequencies, Light is in the middle of the wavelength and frequency, and ultraviolet and x-rays have short wavelengths and high frequencies. Satellites use all parts of the electromagnetic spectrum when they use sensors for different types of data collection.

Systems that use the sun’s light to reflect off surfaces (Passive) use visible, infrared, or microwave portions of the electromagnetic spectrum. While these sensors can provide scientists with a lot of information about smoke, atmospheric chemicals, ocean surface temperatures, and land vegetation, they cannot move through thick, dense covers like heavy clouds and thick forests.

Active sensors also mostly operate using the microwave band so that the waves can pass through clouds. They are useful for collecting data in all kinds of weather. These kinds of sensors are used for looking at waves, rain, forest and city structures, and ice—things you want to see the topography of.

Watch this video on remote sensing, which covers much of what is discussed above

Your Task for Part 1:

Complete this activity from the United States Geological Survey for “Tracking Changes Over Time.” The activity you will do is found at this link, and we also have the pages for it below. We had to cut into two files to load them onto this web platform. If you download from this link, it will be all one document. There are seven questions to answer in this document.

Tracking Change Over Time Understanding Remote Sensing (student guide) _pgs12

Tracking Change Over Time Understanding Remote Sensing (student guide)_pgs34

Make sure to keep a copy for you and upload this to the contest portal. You may need to scan and save as a PDF to upload.

Part 2 – How are computers programmed to recognize objects?

Now that you have a satellite and some sensors to collect data – How do you program the computers that controls the sensors and how does the data from the sensors get sent to Earth and analyzed?There are four Steps or Stages to program satellites to collect data and send it back to Earth where is can be analyzed to recognize objects, like the tracking tags of sea turtles.STEP 1: Program the onboard computer to collect the data and move the satellite

STEP 2: Collect the data and store it on the satellite until it gets sent to Earth

STEP 3: Send the Data Back to Earth and then from the collection station on Earth to the scientists or people who will analyze it.

STEP 4: Analyze the data. Send new programs back up to the Satellite if adjustments are needed.

For more on these steps see this resource from NASA, and also the video below

In the first step, programs in the form of radio waves are sent from a Telemetry or ground station to the satellite in a process called an “Uplink.” The radio waves are GigaHertz or long waves with wavelengths in the mm to cm range. These radio waves can penetrate clouds and get through most atmospheric conditions. The signals must be sent from a ground or telemetry station that can turn the program into radio wave signals. The ground station has to be in the “footprint” of the satellite, which is the area it can send and receive signals from.

Image of the Goes Satellite with the antenna and solar power source labeled. Image from NASA

One advantage of using electromagnetic energy waves is that they are absorbed or reflected by things like clouds, water vapor, and carbon dioxide. Microwave energy can travel through clouds, and this is used to collect weather and communication data from satellites. The main source of the electromagnetic waves observed by satellites is the energy produced by the sun. The sun’s energy is reflected back from the Earth based on the roughness of the surface or the roughness of the features in the atmosphere. The ability of a surface to reflect light is called Albedo, and it is a measure of the reflectiveness of the surface. Snow is very reflective and reflects back 90% of the light. Every material has a unique ability to reflect, absorb, or transmit energy, and these wavelengths are also unique and create a kind of fingerprint of the material. The better the sensor is, the more spectral or light resolution it has to detect different wavelengths.

Sensors

When Satellites use sensors to record the wavelength of the signal, they either use the sun as the source (Passive sensors) or have their own source of light (Active sensors). Both measure the reflection of the wave off the surface of the Earth and back to the satellite.

**Diagram of a passive sensor versus an active sensor. Image Credit: NASA Applied Sciences Remote Sensing Training Program.**

**https://www.physics.udel.edu/~watson/scen103/projects/99s/satellites/howtheywork.html**

Once a program is made and sent to the satellite, it has to be stored by the satellite, so there is a small hard drive on the satellite, much like a thumb or jump drive you may use in school. There is not a lot of memory on these devices, and they have to record and store all programs and data from the satellite. The satellite will send the data it has collected each day down to the telemetry or ground station using radio waves. This is called a “DownLink” and is like a download to your computer. The data has to go to a ground station, and this can only happen when the ground station is in the footprint of the satellite, which can happen once a day or a few times a day, depending on the orbit.

How do programs recognize objects?
The programs sent through the radio waves have to be in binary code that computers recognize and run. Binary code is a series of 1’s and 0’s that are based on a binary number system. Images are made of pixels, which you saw in a previous module, which are also 1’s and 0’s. To get color, though, you have to do more work to convert it to something that computers can read. For more information on how this works, you can visit this website from Wellesley College. The human eye has rods that are light-sensitive and can read red, green, and blue colors. All colors we see are additions of these three colors.

To see more on this, try this slider from Wellesley College, which shows how colors combine to make new colors. Once the image is taken, it gets the pixels converted over to either binary or something called hexadecimal which computers can also read. To learn more about this, watch these two videos

Binary color

Hexadecimal color

Your Task for Part 2:

Complete this activity from NASA Deep Space Communications. The activity you will do is found at this link, but this is a whole lesson, and we just want you to do the two pages below. Your team should print out and fill out each page. There are seven color blanks not filled out in the top table on the document. You will read how to do the conversions and then fill out the table. You must show the teacher your calculations for the math you need to convert BinaryConversions_508 and HexadecimalConversions_508. Make sure to keep a copy for you and also upload these into the contest portal. You may need to scan and save as a PDF to upload.
Color in the mystery picture (hexadecimal) OR the paint-by-number picture (binary) and turn it in. Mystery_PicturesPaint OR ByNumber. Make sure to keep a copy for yourself and upload this to the contest portal. You may need to scan and save it as a PDF to upload.

Part 3 – How does artificial intelligence help computers recognize objects from space?

We now know how computers send and receive images and put them all back together again back on Earth. Right now humans have to do all this work, where they receive a satellite downlink and then have to run it through a processing to convert from the binary or hexadecimal data from the radio waves back over to an image.Sometimes there is interference or noise in the data since there can be clouds or other types of interference from space or Earth radio emissions. There is also so so much data being produced now that we have so many satellites for so many purposes, that this is a big job for humans to do.

Here comes AI to do the work for us

Watch this video from NASA on how AI can help with this job