Directrice, Services corporatifs – Industrie Spatiale
Harjit Sheera est responsable du développement des affaires au sein CGI pour l’industrie spatiale, qui emploie plus de 1 000 professionnels du secteur dans le monde entier.
La série de balados Perspectives on Space proposée par CGI vous invite à explorer l’industrie spatiale avec des experts du monde entier. Dans ce premier épisode, nous expliquons pourquoi il est essentiel pour les chefs d’entreprise de comprendre comment les données d'observation de la Terre sont collectées et utilisées, quel que soit leur secteur d'activité. Les experts de CGI Harjit Sheera, Ron Valk et Auke Blokland discutent de l’impact de l’évolution de la technologie d'observation de la Terre sur le développement durable et sur de nombreux autres secteurs tels que les assurances et l’énergie.
Les données d’observation de la Terre sont collectées à l'aide de technologies de télédétection pour surveiller les terres, l’eau et l’atmosphère. Google Maps utilise notamment ce type de données satellitaires. Au cours des dix dernières années, l’augmentation du nombre de satellites et l’amélioration des méthodes de capture des données, y compris la télédétection infrarouge et radar, ont permis une surveillance de plus en plus précise et le développement d’utilisations plus pratiques.
Selon Auke Blokland, scientifique de données chez CGI, aux Pays-Bas, la qualité des données d’observation de la Terre s’est considérablement améliorée au cours de la dernière décennie en raison des progrès réalisés dans trois domaines :
Ron Valk, qui dirige les services de données spatiales pour CGI, aux Pays-Bas, partage ce point de vue. « Au début, les données n’étaient disponibles que kilomètre par kilomètre. Aujourd’hui, nous atteignons une résolution spatiale de 30 centimètres. » Bien que la résolution et la qualité des informations soient beaucoup plus élevées, monsieur Valk tient à préciser qu’« il faut une plus grande capacité de traitement pour traiter les données. »
Si le traitement de grandes quantités de données peut être résolu en ajoutant de la puissance de traitement, une solution plus efficace consiste à développer des modèles pour réduire la puissance de traitement nécessaire. Par exemple, Auke Blokland suggère de se concentrer sur la détection des changements ou des variations qui présentent un intérêt, pour lesquels un traitement plus poussé peut ensuite être effectué.
La répétition du travail déjà effectué est un autre défi de la gestion des données d’observation de la Terre. Ron Valk observe que les gens téléchargent, collectent ou traitent des données qui sont déjà largement disponibles. Il prévoit qu’en développant et en utilisant des blocs de construction normalisés et en les reliant, comme le fait actuellement CGI, les projets pourront démarrer plus rapidement et les solutions seront plus fiables.
L’amélioration continue de la qualité et de la résolution des données a permis de créer des applications et des services pratiques pour de nombreux secteurs. Auke Blokland explique que par le passé, seuls les changements à grande échelle comme la déforestation et la fonte des calottes glaciaires pouvaient être surveillés. Aujourd’hui, une meilleure résolution et des vols de satellites plus fréquents permettent de mettre en place des services plus poussés.
Ron Valk donne quelques exemples, dont celui d’une compagnie d’assurance qui surveille le bon fonctionnement d’une serre à l’aide de satellites. « Au millimètre près, nous pouvons mesurer les changements et nous pouvons prévenir les dommages en avertissant le propriétaire. » Il est également possible de mesurer les changements thermiques dans les maisons et les locaux commerciaux pour détecter les déperditions énergétiques ou les chantiers non conformes.
Ces mêmes applications qui mesurent la dégradation des serres peuvent également être utilisées pour surveiller le bon fonctionnement des oléoducs ou des ponts. En fait, toute une gamme d’actifs peut être surveillée à l’aide de cette même technique, en fonction des besoins.
Harjit Sheera, directrice, Services corporatifs - Industrie spatiale, chez CGI, prévoit que les enjeux environnementaux pourront être mieux gérés grâce à une surveillance par observation de la Terre des feux de forêt, des inondations, des déversements de pétrole, etc. « Le volume des données satellitaires d’observation de la Terre ne peut que croître, et il est clair que leur utilisation présente des avantages majeurs, notamment pour traiter les questions environnementales urgentes auxquelles nous sommes exposés aujourd’hui. »
Pour en savoir plus sur les nouvelles solutions de surveillance environnementale depuis l’espace, écoutez le balado ou lisez sa transcription (en anglais).
Hi, and welcome to CGI’s series of podcasts, where we will be exploring the space industry in conversation with CGI’s space domain experts. Our experts have a wealth of knowledge across the space industry, and I'm delighted to share this with our listeners today. In this first series, we will be looking at how space data is used across the different industries.
I'm Harjit Sheera, director of corporate services in space. My role is business development across CGI’s global space business, and today I will be exploring Earth observation and the use of space data. Joining me today is Ron Valk. Ron is responsible for space data services within CGI Netherlands. Also, with me today is Auke Blokland, a data scientist in the space data services team, also within the CGI Netherlands business unit. Hi, Ron.
Ron Valk: Hello.
Auke Blokland: Hi.
Sheera: Thanks for joining me today. So, Earth observation is essentially the use of remote sensing technologies to monitor land, waters and the atmosphere. There are obvious advantages when these technologies are satellite-based, but can you describe the technologies that are used?
Blokland: Yes, of course. Earth observation, generally known by the public as things like Google Maps, where you can just look at the map as it's shown. But actually, Earth observation is a lot more. You can also think about infrared, which is in your remote control, but also, radar. All these little parts of remote sensing are actually coming together when we start working with remote sensing. There are different resolutions, different spatial resolutions, different temporal resolutions, so how often does the satellite actually come by and provide us with information? Then, of course, the radiometric resolution, like how dense is the information provided from the satellites?
Sheera: Can you expand on the tradeoff between the different remote sensing technologies?
Blokland: Yes, so what happens, for example, during natural disasters is that the optical satellites are blocked because of clouds. Imagine there is a hurricane, then there are a lot of clouds. The optical satellites don't see anything except clouds, and the radar satellites, the radar actually penetrates the clouds. We can still get an image on the ground surface of what is happening there. We can still see with radar data if there are houses which are completely ruined, or maybe we want to find an airfield, where not a lot of damage is taken. We can actually send people there with first responders to help people who are affected by a hurricane.
Sheera: What kind of data can you get from the methods that we use?
Blokland: The types of data are mainly divided in two parts. You have the optical data, which is like a passive sensor. It just detects what is reflecting back from the Earth's surface. The second part are the active sensors, and active sensors actually send a signal to Earth, and then detect the reflection. Those are two types of satellites which are now in orbit, and both of them are really useful, depending on the situation.
Sheera: What are the benefits, though, of using all of the data from all of the various Earth observation satellites that you've just described?
Blokland: Actually, the combining of data is really useful when we're going to work on a solution, because the one has more benefit than the other in certain kinds of situations. But in general, satellite data is really useful because satellites are orbiting the Earth like 24/7, 365 days a year, so you don't really need someone to go somewhere to catch an image, or fly a drone, or fly a helicopter, fly a plane — that isn't really necessary. The satellites are already there and taking in information every moment when they fly by. Also, what is really nice is that satellites are monitoring really consistently, so every shot is taken with the same characteristics. When you go out in a field and you take a photo with your camera, then you can change the ISO values and such, and with space data, that is all consistent. Every time you get an image, you can compare it easily to the previous one and to the next one.
Sheera: Okay, so how is the quality of the space data that you're using? How has that changed say over the last 10 years or since you've been in the field?
Blokland: What you can see is that there's a trend in smaller and smaller exercises, so the thing is that in the previous decade, the satellites were really coarse, and you could only see the bigger areas, and not much detail. But with more and more satellites in space, you get better and better resolutions. Also, you see that it becomes really cost effective, so it becomes cheaper. With SpaceX, you get cheaper launches, so the satellites also provide cheaper data. What also is really, really easy is that, what I mentioned before, the historical data is always available. You can also look back, see the results, and compare that with the new satellites, the new images taken in.
Valk: Initially, data was available one square only, one kilometer by one kilometer. Now we are up to 30 centimeter spatial resolution, so it's widely available. The resolution, the quality of the information is much higher. Of course, there's a trend that you need more processing capacity in order to process the data.
Sheera: Yeah, that was going to be my next question. I can imagine that with all of the greater capacities and the more processing that you're doing, you must have some additional challenges, as well. Can you describe some of those challenges that the market faces in trying to improve the data quality and the resolutions that you have?
Blokland: Yeah, so the problems we're facing are also tackled by companies like Microsoft and Google. They provide a lot of cloud computing solutions. Some of the problems we face with processing large amounts of data can be solved by just putting in more processing power, but on the other hand, we also try to develop models to reduce the amount of processing power needed. You could just simply process it all, but you could also detect changes. Then on the places where there are changes, which are interesting, do more processing there. So, there are two ways of tackling this problem.
Valk: Another thing, which is also I think relevant, is that we've seen a lot of projects that people are starting from the beginning, so more or less reinventing the wheel. They are, for example, downloading data, or collecting data, or doing some pre-processing, like cloud removals, and all that kind of stuff. We see a trend and an advantage, too, by using standardized building blocks we developed, and linking them together into … a service, so we are able to start faster, and make our solutions more reliable. Because the building blocks themselves have already proven themselves, they're efficient in their processing. They are helpful in our development, so we have one challenge: a lot of projects start with the basics, which are widely available already.
Sheera: The volume of data available is ever increasing, as are the processing and manipulation methods. In your opinion, are organizations able to make the most of what's available and what can be done with the data today?
Valk: Yeah, I think because a lot of our clients are doing this, and what I also see with other colleague companies, is that there are a lot of concept proof values and that the advantage of satellite data is being proven. But the next step to that then, our business or our clients identify — okay, we also need to have an organizational change, because in the procedures it’s described that the measurements need to be done in a certain way. Of course, using space data will impact that way of working. I think that's something that I see with some organizations that they sometimes forget to think about it. Okay, what comes after the successful proof of concept or proof of value? Even if the business case is positive, it still sometimes ends with those struggles, which is those challenges.
Sheera: How do you see these trends translating into applications and services?
Valk: There are multiple potential solutions that are widely available. Because, for example, in the Netherlands, we have an insurance company who is monitoring greenhouses using satellites on their deformation. On a millimeter level, we can measure the changes of the greenhouse, and we can prevent damage by notifying the greenhouse owner: okay, your greenhouse is deforming, and please be aware before the glass in the greenhouse breaks, you have to take some measurements. That's something. And then another thing is measuring thermal changes. Identifying if there are some energy leakages of houses, for example. You can measure using space data on a very detailed level the temporal changes of, for example, houses, also business locations, and to identify if there are illegal activities within those constructions. Those are some examples of how space data is being used, and also to identify crop fields, that's also widely used. But, of course, in the initial days, space data was used for, and there was quite a focus on climate asset management, of course, but initially it was used for large-scale monitoring.
Sheera: Going forward, what sort of trends can you see with the use of these applications and services? How do you think the services will develop as space data availability or quality will change going forward?
Blokland: The big thing I think will change is that, in the past, you could only monitor large-scale changes, so think the climate changes, big scale deforestation, melting of the ice caps, that are the large-scale things you could monitor. Now, with improvement in satellites, you get better resolutions, you get better temporal resolution, so satellites fly by more often. That trend actually helps us to develop services which are more detailed. You could go from large-scale forest solutions to downscale it towards one greenhouse or one building you can monitor. That really helps us in developing more and more solutions, which people didn't even think about. Think of the subsidence, we can now measure a millimeter subsidence if buildings, or bridges, or whatever, are deforming. Well, the thermal monitoring, previously, it was pixels of about 500 meters. We can now bring it down all the way to 50 centimeters, so you can actually see heat changes or differences on one rooftop. These possibilities really, really help in finding new clients, and further developing the presence of space data.
Valk: Also, using this data, it can automate for people, less attractive work, of course, because we do it in a consistent way. You can measure consistently and you're not dependent on the person's opinion. That's, I think, an advantage of using space data. Another thing is the trend that people are becoming older and older. There is a lack of resources for people who want to do some more standardized inspections, or also, impacted by COVID, there was less travel allowed. Using space data, you can also monitor remote areas. For example, we are based in the Netherlands; we can monitor from the Netherlands, we can monitor assets around the globe. That's another advantage and a trend, which I see coming from our customers, that they want to see something which is not currently physically available.
Sheera: Thanks for that, Ron and Auke, you've discussed some very interesting solutions that address some very specific requirements, such as detecting the thermal variations and the subsidence in greenhouses. But I would be interested to know how easily the same applications or solutions can be applied to more than one industry, this question is to the both of you.
Valk: Yeah, that's what we try to do with our CGI EnvironmentMonitor 360 (solution), that's based on building blocks, and connecting those building blocks, we can define all kinds of services. These services can be applied to multiple use cases or multiple clients. For example, as Auke mentioned, the deformation of the greenhouses, the same, of course, applies for monitoring oil production sites. Pipelines can be monitored, but also bridges, as Auke also referred to. All kinds of assets can be monitored using the same technique, and depending on the customer, it can be or needs to have some adjustment. But in principle, the techniques are the same, the same for our thermal service that can be used for monitoring, technical installations — is there a temperature change? Also, is there some leakage from heat pipes? All kinds of use cases can be applied to the same type of services.
Sheera: Auke and Ron, thank you so much for your contribution today. Your experience across the industry has provided us all with a lot of very useful information, and thank you very much.
Blokland: Yes, thank you, and thank you for giving me a platform to promote Earth observation data.
Ron: Thank you for your interest in remote sensing and looking forward to speaking to you soon.
Sheera: Thanks, Ron. Thanks, Auke. We've learned that satellite-derived Earth observation data has evolved significantly over the last few years, with more satellites and improved methods of data capture allowing monitoring to centimeter accuracy. This accuracy, along with the processing techniques, enable applications and services to be developed that meet more stringent requirements. Where previously only large-scale monitoring was possible, we can now monitor individual buildings and temperature changes within specific zones of these buildings.
Our environmental challenges can be better addressed through monitoring wildfires, oil spills, and making efficiencies in labor and costs. However, this also comes with challenges, with increased volumes of data requiring more processing power, and efficient methods for processing. In addition, many organizations, such as utility companies, are not always able to take full advantage of what can be done with the data today, due to evolutions required in their processes and infrastructure for handling this data. In any case, the volume of satellite-derived Earth observation data will only increase, and it's clear that there are major advantages in its use, such as addressing the urgent environmental aspects we are facing today. We hope that this podcast and the views of our experts have given you some insights into the market, thank you everyone for listening.