That’s where machine learning comes in!

We use this fancy technology called “deep learning” to teach computers how to look at satellite images and figure out whether or not there’s a flood going on. It works kind of like magic, but with math instead of wands (sorry if that disappoints you).

Here’s an example: let’s say we have this image here (see below) it looks pretty normal at first glance, right? But if you look closely, you can see some areas where the water level is higher than usual. That could be a sign of flooding!

Now let’s say we have another image that looks like this:

This one is a little harder to tell if there’s flooding or not, but our machine learning model can help us out! It will look at all the different pixels in the image and try to figure out which ones are water (because those might be flood areas).

The way it does this is by using something called a “convolutional neural network” basically, it’s like a really fancy filter that can detect patterns in images. It looks at all the different parts of the image and tries to figure out which ones are similar to other parts (like water or trees).

Once our model has figured out what’s going on in the image, we can use it to predict whether there’s a flood happening somewhere else like in this example:

This one is pretty clear cut you can see that there’s water covering most of the land, so it’s definitely a flood! But what about this image?

Hmm…it looks like there might be some flooding going on here too, but it’s not as obvious as the last one. Our model will help us figure out which areas are most likely to have water (like around those trees), and then we can send someone out to check it out in person!

So that’s how “Satellite Flood Detection using Machine Learning” works pretty cool, right?