Home Blogs Why do you need to explore satellite technologies?

Why do you need to explore satellite technologies?

Let’s say you want to select a new production site, or analyze a new farm field looking to optimize investments. Perhaps you want to identify demographic changing patterns, the progress level of a real estate project, or monitor traffic at a certain port. What about monitoring the effects of wild fire, or preventing the next flood? Everything while sitting at your desk.

Thanks to remote sensing technologies, organizations can now explore new territories and make all kind of strategic decisions like never before. With the availability of worldwide free satellite imagery, nearly all industries will be affected by these technologies. From retail to agriculture, both public and private sectors, we’re on the verge of a serious revolution and anyone who wants to benefit from it should embrace this opportunity as soon as possible.

One technology for multiple uses

In the year 1972 NASA launched its first satellite into space with the intention to understand and manage the resources needed for human survival such as food, water and vegetation. This mission was named “Landsat”, and today every person with Internet can access to more than 40 years of continuous data from all around the globe for free.

After the Landsat mission (which records the longest continuous view of Earth from space, currently a Guinness record), several other providers started to offer all kinds of remote sensing imagery: high or medium definition, multispectral or hyperspectral, 5 or 100 meter resolution, 2 days or 15 days of revisiting time, you name it. The earth observation industry (usually known as EO) is booming and developing very specific products, and several organizations in different industries are already taking advantage of remote sensing imagery:

  • Companies in the financial sector can monitor and estimate early stage activities (tracking inventory levels, assets conditions, light consumption) to anticipate economic trends.
  • Organizations in the fishing industry are using satellite imagery to identify the most productive areas, monitor red tides and measure ocean temperature.
  • Firms like Orbital Insight claim they can track the health of major retailers by analyzing images of their parking lots (using it as a proxy for occupancy rate).
  • Players in the energy industry can use remote imagery to monitor assets, measure environmental impact and screen competitors.
  • GlobalXplorer is an online platform that uses the power of the crowd to analyze satellite images currently available to archaeologists.
  • Maritime surveillance entities can use satellite imagery for sea border activities, monitor maritime traffic and detect illegal fishing.
  • Organizations in the forestry sector can classify land cover, track changes and identify degradation areas.

satellite imageryAutomated algorithms detect and count cars in satellite imagery. Source: Orbital Insight

These technologies have also become extremely important for crisis management. Think about droughts, flooding, fires, snowing, or any other natural disaster. All these events can be predicted and monitored using satellite imagery. How many lives could be saved with this technology?

Glaciers can be monitored for climate change purposes, and even contamination sources like oil spills in water can be easily identified with satellite imagery, which represents a priceless capability for humanitarian purposes.

Remote sensing technologies are a great example of technology leveraging human capabilities since they provide information far beyond what we’re able to perceive and analyze. They give us the opportunity to see light in broader wavelengths than our eyes are capable of, and combine different spectral bands to detect objects or phenomena that would be invisible to the human eye. Can you imagine having the ability to detect the moisture content of soil, spot temperature variations in a specific area or assess plant vigor with your eyes? This is possible by using satellite imagery.

Agriculture is first in class

Probably one of the best matches between remote sensing imagery and an industry is now taking place in the agribusiness sector, where artificial intelligence and robotics are already playing a crucial role.

By performing algebra with spectral bands, vegetation reflectance properties are used to calculate vegetation indices that help farmers, governments and investors to take better decisions. The main idea behind this is that healthy vegetation reflects light differently than weak vegetation, which allows us to identify events even before they become visible to the human eye. Through remote sensing technologies it’s possible to digitally intensify or decrease different vegetation properties, which enable us to estimate coverage, growth and production rates, vigor and biomass, among other calculations.

Can you imagine what would be the opportunities for someone who could estimate crop productions in a region, a country or the entire world? How would that information affect real economies and financial markets?

The company Descartes Labs states it can beat the accuracy of the US Department of Agriculture (USDA) predictions, and their algorithms are getting even more precise as they keep learning from data. They believe their solution is a game-changer in crop prediction, and they are already tracking crops in different continents. We’ll have to wait and see if they are right.

nasaOn the left, a false-color Landsat image captured over Grand Forks, N.D. August 30, 2005 to evaluate the health of crop fields. The image on the right shows the same land area in true color as the natural eye would see. Source: NASA

Remote sensing imagery is perfect for measuring plant population density and size to determine their evolutionary development. Diagnosing plant disease is essentially a pattern recognition challenge, and if farmers detect deviations during the growing stage they can correct them before suffering yield losses.

This technology can also be used to predict and monitor crop diseases and pests to balance the application of agrochemicals, hence reducing costs and residues. But the industry has gone a step further, since it combined these tools with digital elevation models built from satellite imagery to generate maps that instruct machinery (like spraying machines) to put the precise quantity of seed, fertilizer, herbicide, or any other input in the area where it’s required. These tools are so important that they have given birth to a new concept in agriculture: precision farming.

The rise of Machine Learning

Data from images is just one side of the equation, and there is a need to provide answers to complex questions such as: how can we predict environmental events, customer behaviors or any other event that doesn’t seem to be predictable? It’s not enough to interpret satellite imagery, and to move a step forward we can turn to the machines for help.

Several companies are boosting imagery with image and pattern recognition algorithms driven by artificial intelligence to reach full capabilities. Just like in the agribusiness sector, remote sensing technologies will merge with artificial intelligence leveraging better solutions for organizations in all industries. This is only a matter of time.

In this sense, one of the most promising fields under the discipline of artificial intelligence is machine learning, which is basically a way for computers to learn from data without being specifically programmed for it. By employing different algorithms (which are set of rules that computers can follow), computers can process data like never before. This technology is extremely valuable to discover patterns and rules in sets of data, which combined with remote sensing imagery can change the way we manage the environment, our business, and even ourselves. Artificial intelligence is blending with remote sensing imagery in different industries to improve efficiency and open new possibilities everywhere. Let me give you some examples:

  • Stanford University is using machine learning to predict poverty by estimating household consumption and assets based on satellite imagery.
  • Two years ago, Digital Globe partnered with Uber to improve the precision of existing maps, and similar uses are also being explored in the autonomous-vehicle industry to develop smarter vehicles.
  • Last year Orbital Insight applied this method to measure Hurricane Harvey in the US with extremely precise results, which comes out as great news for flood models, impacting governments and insurance companies alike.
  • Another big tech player like Microsoft is already offering machine learning solutions applied to satellite imagery in its Azure platform.

Machine Learning is perfect for image recognition and classification, and the good news is that it’s just getting better and better at it. If you want to automatically recognize objects, cluster an image by similarity or classify what you see in it, then this is the right technology.

Using computer algorithms that resemble our biological neural network functions, it’s possible to identify patterns in images independently from human effort or prior knowledge in order to classify these images based on their characteristics. Models can also be trained with millions of images or external cross referenced data to improve its accuracy far beyond our human capacity, which makes it very difficult to imagine where the next frontier will be. 

How does the future look like?

Although historically the satellite industry was characterized by public players dealing with long and multibillion dollar projects, the advances and democratization of technologies helped to drop prices dramatically and opened opportunities for commercial and private players as well.

It’s calculated that over 2.500 satellites are orbiting Earth providing all kind of services and data, and it’s almost certain that the number will keep increasing.

Last year MDA announced the acquisition of the company Digital Globe for USD 2.4 billion, creating Maxar Technologies and becoming one of the biggest space technology firms worldwide. Also last year, Planet acquired Google’s Terra Bella business and sent 88 satellites aboard a single rocket (the largest fleet of satellites launched in history).

There’s no doubt a significant increase in available data is going to happen very soon, and it’s highly possible that basic imagery solutions will be bought as commodities in the very near future.

Another rising industry is the nanosatellites one, which will drop the costs of imagery even further. Fast changing technology is allowing to produce smaller satellites with similar capabilities as the conventional ones, at a fraction of the cost. At this pace, every one of us will have deep access to space, and this will transform almost every aspect of our lives.

Remote sensing information has a very important characteristic: it’s pretty much based on open-source technologies and platforms. NASA and other organizations provide all kinds of different satellite imagery and models for free. QGIS is a free and open sourced platform that enables to create, analyze and publish all kind of geospatial information. And the list just goes on and on.

In this scenario it’s possible to achieve big endeavors literally with nothing more than a computer and internet access. These conditions have rarely been presented in our history before.

No matter in which industry you’re working, my advice is that you explore remote sensing technologies and look for opportunities for your organization. I’m sure you’ll be surprised by the results.

Also Read

The tremendous potential of Machine Learning in satellite imagery