Tackling environmental challenges through satellite imagery and artificial intelligence

Published on 2022-07-08

Startup Story

Listen the full podcast or read the article below!

Remi Charpentier told us how, with his startup Tesselo, he is trying to address the problems associated with the lack of a dedicated satellite imagery market for businesses.


About Tesselo

The other co-founder and I started thinking about Tesselo in 2017 when Portugal was devastated by a huge fire: more than 200,000 acres of forest burned in a couple of months.

It is very interesting to know that only 20% of the forests in Portugal are more than 20 years old, while all the other 80% are young forests that were planted earlier because they burned in the past.

This problem shocked us a lot and prompted us to take action on climate change. In addition, the European Space Agency launched the new Sentinel satellite system that offers many new possibilities for earth observation and working on fires.

So we decided to take advantage of these new satellite images and use our expertise in artificial intelligence modelling to make sense of these new images and apply them to solving other environmental problems. 

Tesselo started from a technological idea that we wanted to apply to solving environmental problems, and of course, we started working with the forestry industry here in Portugal.

Tesselo's technology: low or high-resolution satellite imagery?

We have one technology that we apply to many different problems. This stems from the fact that satellite imagery today is not suitable for environmental planning and solution. 

Indeed, on the one hand, there are free satellite images - which everyone can access without paying - and these data are excellent, but they have a low resolution that doesn't allow you to see many things in the image, so you can't build a really real application with this kind of resolution.

On the other hand, there are high-resolution commercial images, such as those from Google Maps, where you can see very small details. The problem is that these images are very expensive, so when you want to cover an entire country and monitor a forest with high-resolution images, it is not economically feasible.

We have developed technology using super-resolution, which basically turns these low-resolution open images into high-resolution images. We are monitoring, for example, vegetation around power lines, which is a big problem because the vegetation can ignite with power lines, and to get results we need a resolution better than 10 meters. But the power lines span an entire country: in Portugal, which is a small country, there are 40,000 kilometres of distribution lines, so we developed a vegetation map with a resolution of one meter around the distribution lines, and that would solve the fire problem.

The unique aspect of our technology is that we use mostly only open data, but we can provide the same accuracy as solutions based on high resolution.

The different stakeholders who can use satellite imagery

We work with four types of customers. 

The first is industrial forestry: paper manufacturers, furniture manufacturers, and any other type of company that plants trees on an industrial scale, usually with a monoculture of the same trees. We have developed many solutions to help them improve forest management and do precision forestry.

The second category of clients we have is carbon-related, essentially reforestation and afforestation projects. We help these companies plant trees, first to find the best place to plant new trees, then to identify species to plant. 
It is important that when you plant a forest it is for 100 years, so when you plant a particular species you have to consider what the climate will be in 10 years and 20 years. For example, it would be risky to plant a pine species in Spain or the south of France today, because it is not certain that the climate will adapt to the pine in 20 years. 

The third category is utilities: we manage and monitor all vegetation risks around critical infrastructure, power lines and oil pipelines, where vegetation management is very important to reduce the risk of floods and fires. 

We also work with public institutions, municipalities, and cities on different projects, typically optimizing landscape planning: how to organize the land to be somehow more sustainable or how to minimize the risk of fires, etc.

Consumer response to satellite imagery and artificial intelligence 

Our solution has a strong impact on customers because most organizations do not use imagery, so any image-based solution is considered new and will have some kind of impact because it will bring new information that people are not used to having. 

We have had very positive feedback on the impact of our solution, for example in the forestry sector in Portugal, they estimated that Tesselo saved customers 25% on-field operations costs because by using our information they were able to reduce the number of field visits needed. 

Similarly, for our customers in infrastructure who need to cut trees that run over power lines, our solution can reduce the operational costs for vegetation management by about 30%. 

So we have an economic impact on our customers' operations that is immediately visible. As far as the environmental impact is concerned, it is an indirect or long-term impact: for example, by reducing utility operating costs, we also reduce the risk of wildfires by making sure that our customers can better manage the vegetation around the power line.  

Example of Tesselo application

I think the greatest number of products and solutions we have developed are those made for the forest industry. 

As a case study, I will show you that Portugal: is the largest paper producer in Europe. In fact, in all of Portugal, there are about 1.5 million hectares of pine and eucalyptus trees, and for them, we developed the forest inventory that every week identifies all the trees in Portugal. We survey the species and calculate a series of indicators such as the age of each tree, the volume of timber in the trees, and the time of last cutting and last burning. 

In this way, every week we have a complete picture of the forest landscape and can see, with great accuracy, what the forest activity is in Portugal on a continuous basis. 

This is our flagship product: it chooses 10 different models and is quite complex.  We are now distributing similar solutions in countries like Spain, South America and Africa. 

The main obstacle of satellite imagery: identify the market

The main challenge we are still facing is not technical, but market: most of the organizations that could benefit from satellite imagery today are not using it because they perceive it as not very useful, too expensive, etc. 
All clients in Portugal have been using paper for the past 150 years and have never had to use satellite imagery, so we need to educate clients and show what the benefits of remote sensing analysis are. 

This is true for the private sector, but for the public sector it is more complicated, and I will take a very clear example of forest wildfires. You might think that fires are very important-everybody cares about burning forests and, as you can see on TV, California, Australia, Portugal, and Italy. 
You might think that someone will pay to reduce this risk, but in reality, there is no one to buy a solution to reduce the risk of fires: the government will not buy it, municipalities will not buy it, and fire departments have no money for it, insurance companies do not insure natural fires, and so on.  
Although forest fires are a dramatic problem, there is no market for solutions to mitigate them.

The only way we have found today to fight fires is to help utilities reduce the risk of fires with infrastructure, but we have not yet found a customer who is going to buy comprehensive fire risk mapping and take landscape planning actions to mitigate this risk. 

There is a great need, but a market is not necessarily behind it. 

The main challenge for us as environmental engineers is to identify who are the stakeholders of the problem we are looking at and who is ready to pay for it

For example, with regard to fires, we provide a solution to utility companies, but they don't care much about fires-they just have to cut down trees, and that costs them a fortune. So, we help them reduce costs, but we also have a direct impact on reducing fire risk.  
In the beginning, we thought we were working with the fire department, instead, we were working with the utility companies, which is fine, but it was not apparent at the start.

Tesselo's future applications

Today we work mostly on forestry-related problems and vegetation management. 

Tomorrow we want to expand our expertise to agriculture and to other kinds of environmental problems like flooding and drought. 

About the author

Remi Charpentier

CEO at Tesselo