TakiWaehere – New Zealand Geospatial Hackathon

[Intro titles: Introducing The AUNTS. Finalists from TakiWaehere: The Geospatial Hackathon]

[Olivier standing in front of Auckland University Mission Operations & Control Centre computer screens]

Olivier: So we were the Auckland University Neuron Transmitting Students, or the AUNTS for short.
[Shot of The AUNTS team members: Amelia, Ben, Sam and Olivier]

We're an Auckland Uni based team with computer scientists and civil engineers for the New Zealand Geospatial Hackathon. So we chose Managing Emergencies.

[Olivier standing in front of Auckland University Mission Operations & Control Centre computer screens]

The reason we chose that challenge is because some of the team members were actually quite involved through their own studies, and those sorts of areas. So looking at things like studying earthquakes, and things like that.

[The AUNTS team sit at table, Olivier writing in notebook, Amelia on a tablet, Ben and Sam on laptops]

So we created a tool that uses satellite data to detect when a flood has happened in New Zealand, where it's flooded, and also detects all the flooded areas.

[Close up of laptop showing the tool]

[Olivier standing in front of Auckland University Mission Operations & Control Centre computer screens]

And then from that was able to estimate the risk to local infrastructure, houses, roads, and so either local emergency help or people in the houses or concerned relatives could estimate the damage, see where the flooded areas were, and just view this all on a live map online.

[Amelia standing in front of Auckland University Mission Operations & Control Centre computer screens]

Amelia: We had four people and each one had a different task.
[The AUNTS walking down a uni hallway]

We had Ollie who did the satellite image processing - which I had no idea about what was going on - but the output went to me and I did the GIS work, meaning overlaying different layers and seeing how they work with each other. That then actually went to Sam who created this web map, which means that all the information that we created, was then accessible for everyone else. And Ben, actually, from the start, he was working on the report and like feeding us information about the background. So we could actually frame the whole project around this one problem.

[Amelia standing in front of Auckland University Mission Operations & Control Centre computer screens]

Our project, or our tool, was addressing a real life problem and a real life solution that is for everyone. So it was for a particular group or particular stakeholder, because our tool can be used by emergency response teams, it can be used by people who are actually affected in another way, or even people who want to find out more about the event itself, without being affected at all. That's why we thought it was really important to create this web map as one of our outputs, so that everyone could access the information.

[The AUNTS team sit at table talking, Olivier writing in notebook, Amelia on a tablet, Ben and Sam on laptops – moves to close up of tool]

Let's say you want to see if you can like still provide transport, you can use the tool to see if certain routes are still available.

[Amelia standing in front of Auckland University Mission Operations & Control Centre computer screens]

If they weren't, you can start looking into alternative routes. Same goes for things like rail, if the rail tracks are flooded, then they can't be used.

[The AUNTS team sit at table talking, Olivier writing in notebook, Amelia on a tablet, Ben and Sam on laptops]

And one feature of our tool was not just to detect the flood that already occurred, but also to monitor floods. So imagine you have all this space data from let's say, New Zealand or even around the world,

[Amelia standing in front of Auckland University Mission Operations & Control Centre computer screens]

The process of actually identifying where floods can occur can be automated, which means that the images can be directly fitted into the tool. And the tool can even alert you if it sees that somewhere, where maybe nobody is around at the moment, where there is like water accumulating, which could potentially be a flood.

[Olivier standing in front of Auckland University Mission Operations & Control Centre computer screens]

Olivier: So Maxar satellite imagery was very good to use. The first, most defining characteristic is the 30 cm resolution, which allowed us to see between houses, like see roads, even like flooded banks, details that just wouldn't be visible with other satellites out there. Another advantage of using Maxar satellites was the multispectral data.

[The AUNTS team sit at table talking, Olivier writing in notebook, Amelia on a tablet, Ben and Sam on laptops]

And that allowed us to do some clever techniques with the near infrared sort of side of the visible spectrum - of the not-visible spectrum I should say –

[Close up of tool]

which allowed us to accurately detect which pixels were flooded in the image and which ones were not, to make the tool possible.

[Olivier standing in front of Auckland University Mission Operations & Control Centre computer screens]

There's lots of future developments that are happening in the world of satellite imagery that are very exciting. The most obvious one is that the resolution of satellites is improving both spatially and temporally, and it's exciting to think that in some decades, we might be able to see these things, these natural hazards happen on a maybe even an hourly basis, clouds permitted, and over the whole with a global coverage. And this seems like it very well could be the future for emergency response. There's also lots of other sorts of things that can be monitored using satellite data. And just on topic of managing emergencies, you can look at things such as earthquakes, detecting those from space, and I mean, or other emergencies such as algal blooms in lakes. These both have, sort of, important real world events, effects, and currently the methods of keeping track of them aren't as effective as they could be, if they were just you know, monitored from space, which is a very achievable task. And people have already started doing this around the world, and it would be good to see New Zealand sort of following in those footsteps.

[End title: New Zealand Space Agency logo, New Zealand Innovative Partnerships logo and Maxar Technologies logo]

[Intro title: Introducing: The Interpolators, Finalists from TakiWaehere: The New Zealand Geospatial Hackathon]

[Morgan sits in front of window]

Morgan: The Interpolators is our team name,

[Team put on hackathon branded t-shirts, Morgan, Hannah, Helen, Ander, Brendan and James]

we all just kind of came together, we had a lot of complementary skills, and that's how I would describe the team. A bit of a, extremely capable hodgepodge.

[James sits in front of window]

James: Yeah, so the challenge that we picked was waterway health. So monitoring the smallest New Zealand waterways, largely, they're not monitored particularly well, just because there's so many, and it's just kind of impossible to be able to do with our manpower. So it was something that we felt quite strongly about. Waterways are really important, because we can kind of extrapolate a lot of information from waterways, about other environments as well.

[Morgan sits in front of window]

Morgan: So during the 24 hour hackathon, we created a prototype tool, and it performed a series of very specific steps on one image of satellite data.

[Close up of the tool on a laptop – showing a satellite image of a waterway]

The tool finds water and it does this by looking at the different wavelengths in the image.

[Morgan sits in front of window]

And there are certain combinations of wavelengths that you say is water. In the same way that if a human looks at something that is like blue, green, clear, we say its water, we can do the same thing with the satellite and some other more sophisticated wavelengths of light.

[Close up of the tool on a laptop – showing water indices]

And then the program went through a series of health indices for the water and for the area near the water.

[Morgan sits in front of window]

So the idea is, in the same way that we could look at an unhealthy river, and it looks brown and gross, and very turbid, perhaps, we can use our satellite data and different wavelengths to compute an index of health, for the specific pixels it sees.

[Team sits at a table talking]

We got the tool to represent the overall health index on a color scale, on a re-output image overlaid on top of the red, green, blue image of the waterway.

[James sits in front of window]

James: One of the really good things about satellite imagery is that it has lots of different layers, which gives us extra information. So usually, with the human eye, you can only see red, green, and blue, you might have heard the word, the term RGB before. And so from that, we can only make out a very small spectrum, section, of the electromagnetic spectrum. With the satellite, we can see eight different bands. So you might have heard terms like infrared and near infrared and things like that, which are included in the satellite. So what we can get from that is all sorts of extra information that the human eye can't see. Which sort of allows us to access a lot more information, a lot more quickly. A lot of the factors of river health are very small, like they're dissolved in the water, very small particles. And so the problem is usually with something like generic satellite data that you might get to see on Google Earth or something like that, it's 30 meter resolution, 15 meter resolution, so every pixel on that image is 30 meters, or 50 meters on the ground, where the Maxar data, 30 centimeters, so everything that's 30 centimeters, you can see a 30 centimeter brick on the ground, you can make that out on the image.

[Close up of Brendan on a laptop and Ander talking to a team member, Morgan and Hannah look at a phone laughing]

[Shot of Ander at a laptop, talking to Hannah. Brendan and Morgan discuss something at the window]

Morgan: Obviously, this tool isn't our final product. We, for example, we want to include all sorts of things. For example, we want to include stuff like machine learning to classify the riparian area. So say, you know, is that a house? Is that a tree?

[Morgan sits in front of window]

This can help our tool be more specific, we want to incorporate a time series aspect. So we made a tool that looked at one image, we want to eventually have a tool that can look at one waterway over time, and detect changes in the health index it calculates. And we also want to really refine our health index.

[Morgan sits at laptop – there is a close up of the tool over Morgan’s shoulder]

James: So businesses could definitely benefit from the tool quite a lot and regional councils as well. A lot of businesses have environmental responsibilities to maintain.

[James sits in front of window]

So you know, construction sectors and farming sectors and things like that, they've got sustainability goals to meet. And like regional councils as well, they're required by the government to monitor their waterways. And at the moment, that's quite costly. And ideally, with the development of this tool, it will hopefully provide them with a more cost effective method to monitoring more waterways as well as monitor them better.

[End title: New Zealand Space Agency logo, New Zealand Innovative Partnerships logo and Maxar Technologies logo]

[Intro title: Introducing: The South Island Students, Finalists from TakiWaehere: The New Zealand Geospatial Hackathon]

[Zach sits in university study room]

Zach: Team South Island Students, we are, by the name, we are all from the South Island down here at the University of Canterbury. We are all mechatronics students.

[Shot of the team sitting at a table on their laptops, Fletcher, Zach, Hank, Henry, Georgia and Sam]

We all come to the project to tackle the changing coastlines, the changing coastlines theme. And what we wanted to achieve with this is, see if Maxar's high resolution satellite imagery could be used to overcome the very large expense of hull and anchorage collisions with shallow water features like rocks or coral. And we hope to overcome this using our tools See Reef, which we developed during the hackathon.

[Zach sits in university study room]

See Reef aims to overcome this by providing near real time information on where the submerged rocks and coral bommies, which can often sometimes move with tides, where they are to a resolution that hasn't ever been achieved before.

[Screen recording of the tool, zooming into satellite imagery of Christchurch coastline]

One of the opportunities at high resolution satellite imagery gives us, is the ability to look at autonomous ship control.

[Zach sits in university study room]

Chart plotting software in the traditional sense, is a bit like a road map,

[Screen recording of the tool, showing the plotting of a ship journey]

It'll tell you how to get to point A from point B, but it's not going to tell you that there's a, there's a pothole 10 meters in front of you.

[Zach sits in university study room]

What we're able to do with See Reef is tell exactly where the potholes are, where the shoulder of the road is broken off and help guide the ship around all of these short, close proximity hurdles to safe passage.

[Fletcher sits in university study room]

Fletcher: So we would take an image of that Maxar provided us,

[Screen recording of the tool, showing the index of bands used on a satellite image]

We would extract out the bands that we were to use, we would then, using those bands mask out the land. Once we've found the mask of the land, we could then basically find all the obstacles in the water.

[Fletcher sits in university study room]

Once we identified them we could then convert them to GPS coordinates and display them on the website. Something, I think that made us really unique was our front end application. So we use a leaflet JavaScript framework to build a map in which all our data could be plotted on it, so it allowed a heat map and that could be overlaid with a ship's potential route, and anchorage points which were pre calculated, and thus could be navigated to.

[Zach sits in university study room]

Zach: The benefit with using Maxar imagery is A, it's really high resolution, we're able to actually identify rocks, which would typically not be identified in say traditional chart plotting software,

[Screen recording of the tool, showing the index of bands used on a satellite image]

But also at the rate at which it's updated. Every few days, we'd be able to get a completely fresh update of what these coastal bays have in both low and high tide situations. Just that alone would be able to be invaluable to both commercial and recreational skippers.

[Fletcher sits in university study room]

Fletcher: Maxar's data was really useful as unlike normal satellite data that you would get from the internet, we have a large array of bands, so head near IR, which was really helpful for identifying the coast and land.

[Screen recording of the tool, showing the index of bands used on a satellite image]

We then also had the coastal band at the other end of the spectrum, which was better for identifying obstacles in the water, which is the kind of thing you can't do with normal satellite imagery.

[Zach sits in university study room]

Zach: See Reef would be great as a standalone app. It can be accessible to both recreational and commercial skippers to see on their phones. But what would really be valuable is the integration into existing chart plotting software such as Navionics. Having the ability to overlay information onto the already very impressive chart plotting software that's out there, we would be able to give a new layer of information for when skippers are coming in for anchorage, or trying to make safe passage through hazardous areas which are so common in New Zealand's coastlines.

[Team sits at table on laptops, talking and laughing]

Fletcher: While we have adapted it for coastlines, it does have other uses. So one could be self-driving cars, in which, because roads change all the time and so the ability to rapidly update current roads for maps and self-driving cars and that it's another data input for them to use.

[Fletcher sits in university study room]

Another application could be around Antarctica for scientific journeys as due to climate change, ice sheets are consistently changing and the coastlines are very variable. So the ability to rapidly update a part of the world that's not often, super well surveyed would be very valuable.

[End title: New Zealand Space Agency logo, New Zealand Innovative Partnerships logo and Maxar Technologies logo]