Contribution of deep learning and data science in satellite image processing.

We are in a country having a pretty good space department, we showed a lot of good work. And we are very proud of them. One of the things they've been doing is to put out satellites and satellite images from our satellites. Of course, there are also commercial companies, which give you satellite images. And there are many good use cases of what you can do with satellite images. I'll give you an example. In China, we always have this water problem. So we have some reservoirs around the city. At some point, you'd like to know how much water is there and so on, of course, the reservoirs have levels and all that, determined to be able to see, but what about collecting data from satellites to see what the extent of the water is, and use that to estimate how much water it actually is. So the satellite has very high resolution cameras, you can actually see the extent of the water. In fact, many of you would have seen Google Maps and  Google Earth, where you can see exactly your rooftop and so on, right? So that is one use case, you can try to track natural bodies, like forests, and water bodies, and so on. Another example could be that you can, in fact, one use case that I know is there was a company that was trying to tell you for your house, how much solar energy can you get? So by looking at your house and Google Earth, they can tell you what the area of your terrace is. And if you were to put solar panels in a particular angle, how much energy will you make in one day, for example, on an average day, so this is a good example of what you can do. One more example could be to model for example, playgrounds. So we have lots of playgrounds in the country, if you want, if I asked you to give me a region in Tamilnadu, that there is no playground within five kilometers. Right? So how will somebody solve this problem today, and they don't have a list and they don't have GPS locations, and they do it. But with satellite imagery, you can think of doing this right. So you can actually say that such things which are around and large are typically playgrounds, and I can detect all the playgrounds in the map, and then try to find locations where there is no playground within five kilometers. And maybe this is an input to the government to try and create a playground. So there are many good use cases of satellite image processing. It is the intersection of problems that are solvable, and also valuable.