Building a scientific UROV

Hey all,

I am building a really exiting new project with my team. We’re planning to develop a UROV with visual recognition for all animals in the Great Barrier Reef. It will have autonomous tracking, it will be tethered directly to a buoy which will have a radio transmitter back to the controller. It will go 530 feet deep and have two articulated claw arms and a sample cage.

It will be able to collect samples of the water for analysis, calculate the number of certain fish in certain areas, identify invasive species.

This is just the beginning. It’s an ambitious project and well need all the help we can get.

Hope you guys are interested!!


@aidenjmccomiskey That sounds great. I have some software I use when out in the forests to identify plants and animals in real time through live video/images of them. It would be great to start using that software for the ROV.

You’ve inspired me into seeing if the software I currently use can be adapted to work with the ROV.


If that works, please let me know, I would be really exited.

I’m glad we were able to inspire you!

1 Like


This coming weekend I’ll see if I can access the ROV camera from the software. If I can, I’ll take the ROV outside and test the software on out plants in my yard that I already know it has identified. If that works, I’ll plan a trip to the bay and see how successful it is to identify plants and animals in the ocean. :smile:


G’day aidenjmcomisky,

I am curious to know more about your project and how you think we might be able to help. If your project is for something more than personal interest then you should look at some of the work being done at QUT.

Yes our project is more than personal. Thanks for the recommendation, I will definitely check them out!

How did the test go? I am really exited to learn more!


@aidenjmccomiskey Before I got the ROV out, I tested how well the software would recognize plants and animals in the ocean, using images. Though the software does great on land plant and animals in the field or from images, it seems to do lousy with sea life. :frowning_face: Learning that, I put off trying to couple the ROV and software.

I plan on doing more testing how well it recognizes images of sea life, before field testing it in the bay. I also plan on contacting the organization that wrote the software to get input from them.

I wish I could have better news. I’ll keep working at it though.


That sucks. Keep me posted, I’m exited to know more. Also, if you wouldn’t mind telling me, what is the name of the company from which you got the code?

1 Like

@aidenjmccomiskey The software is not code, its Seek, an app by iNaturalist. I have it installed on a tablet/laptop that I use to push live footage from the ROV to the internet. Seek can process live video or stored images. Since I can select the video from the ROV as the video source for Seek, it will be able to process anything being seen by the ROV. The other option would be to process the images you pull from the ROV’s video later, and feed them to Seek one image at a time. either way, you get the same results.

I’ll do a little demo video when I get a chance and post the link to it on the forum.


1 Like

Where are you based?

If you plan to travel to Brisbane to meet with Prof. Dunbabin and his team then we could catch up for a chat. Some of what you described is of interest to me and I am happy to collaborate if the terms of your project allow it.