For my masters, me and another engineer were tasked with developing an ROV using the T500 thrusters that could be used for habitat mapping. A downward facing ILX-LR1 camera was integrated with the ROV to take high resolution photos for photogrammetry and the camera was fitted with a special lens to take multispectral images to showcase the capability of using the ROV to identify different aquatic plants in a shallow water environment.
The Imaging and Analysis for Seafloor Classification (IASC) ROV had a weight of 30.63 kg with a length of 550 mm, a width of 812 mm and height of 500 mm. The video attached below shows the integration of the ILX-LR1 with a BlueROV2 between 00:00 - 2:52 and the development of the IASC ROV between 2:52 - 7:46. The ROV was tested in a dive tank before being tested in a freshwater environment at a depth of 10 meters.
We found that the IASC ROV outperforms the retrofitted BlueROV2 in maneuverability and stability, making it ideal for imagery collection. We also only ran the IASC ROV at low gains using the 14.8V 18Ah battery and even at low gains it was very capable of navigating with the T500 thrusters in the environment it was tested in. The downward-facing camera that is mounted on the IASC ROV also gave a unique perspective for navigation having two different directional views during flight and would make it much easier to collect imagery of the seafloor. We made sure there was room for expansion with mounting holes for additional attachments e.g (grippers, water linked DVL) and a additional enclosure for more control software or electrical components needed.
We would love to know people’s thoughts on the overall project and the design, what could be improved or revised on the ROV to make it better if a similar project was to be attempted e.g (redesigning the ROV replacing the cylindrical enclosures for the new watertight boxes with a heat sink for cooler electronics and a more compact ROV.) or what could be integrated or expanded on with the existing ROV to make it more functional and capable in the future e.g (upgrading to the 22.2V 28Ah battery or top side power, adding the water linked DVL to allow measurement of distance from sea floor or for a dead reckoning system.)
This is super cool build, nicely done, @Lsbutler21!
We’re doing somewhat similar work with BlueROV2s to conduct photo/video surveys within kelp forests along coastal Washington (see here). I would think that ROVs with the T500 thursters could be really useful in areas of the coast where ripping current challenges surveys.
Regarding your next steps: supplying power over the wire was a game-changer for us. We use the Outland system and x2 200Ah lithium-ion batteries on our vessel. Having a DVL is also quite useful for our ROV surveys, as it, e.g., allows us to lock-in any given altitude above the seafloor to improve the consistency of our survey imagery, e.g., via surftrak, which is now a part of ArduSub 4.5 Beta (and also see surftrak_fixit in the BlueOS Extensions for surftrak setup / monitoring assistance).
I do have a question . . . we were/are interested in the Sony ILX-LRI camera you used . . . how did you integrate it with the ROV? A colleague of ours attempted the LAN connection, but reported some issues that I don’t believe were resolved. I’m aware there’s also a local wifi connection option as well, but we opted not to pursue that. How did you go about the ILX-LRI integration, and what was your experience??
HI @zhrandell -
Surftrak is on version 4.5 stable now!
Love the fixit extension - hadn’t seen that!
I’m eager to hear about the Sony camera as well, particularly if it powers on with the vehicle and doesn’t require opening a housing to push a power button…
Hi
The video looks cool.
Just wanted to know what wavelength of light was used to map the seafloor plants using the multispectral camera.
I have a MAPIR camera and would like to know if this can be effective in doing a sea floor mapping as well.
Thank you
Thanks for the response to the post, we were able to connect the ILX-LR1 topside allowing for remote control of the camera using a LAN connection to connect the camera to the computer with imaging edge desktop being the software used for remote control. We tried this first with just the ILX-LR1 connected straight to the computer and then once we verified that we could connect it and remotely operate the camera we connected the camera to the ethernet switch on the ROV and tested it again for communication.
The Sony ILX-LR1 camera has HDMI and USB-C ports. A USB-C to Ethernet adapter was used for communication with the PC. The camera and PC were set to a static IP address, the IP address for the ILX-LR1 is set to communicate with the topside computer allowing control of the camera through the Imaging Edge software. The Raspberry Pi is set to 192.168.2.2 so we set the Topside Computer to 192.168.2.1 and the ILX-LR1 to 192.168.2.10 . The images taken by the ILX-LR1 camera can be transferred to the topside computer and saved in folders through the LAN connection, which also gives a live video feed on the Imaging Edge software.
I hope this response answers your question.
Kind regards,
Liam
the six filters are represented in post processing as 14 individual wavelengths, which can be overlaid using the software to identify different analyzing metrics
Filter 1 - 405 | 570 | 710 nm
Filter 2 - 430 | 550 | 650 nm
Filter 3 - 450 | 560 | 685 nm
Filter 4 - 850 nm
Filter 5 - 525 | 630 nm
Filter 6 - 490 | 735 nm
the few results from our project shown in the video link at 2:44 show that by using the multispectral lens it is possible to get results that can distinguish areas of greater vegetation. However I don’t know if it can be said that this can accurately map the sea floor to identify different species or if a MAPIR camera would be effective in doing so as I’m not well versed in the area unfortunately.
What sort of latency are you achieving with this setup? I am running a Canon R6 mk2 plugged into a mini stick PC, which runs the Canon software. This is then in turn connected to the ethernet switch and can then dial in over Remote Desktop. I have found this method to give very low latency compared to using VirtualHere. Also, what USB-C to Ethernet adaptor do you use, please?
I couldn’t give you a measurement on the latency unfortunately Marcus but we didn’t have low latency problems, it can kind of be seen in the video at 5:54 when we’re taking photos the pause in the video is the camera taking the photo. The USB-C to Ethernet adapter used was a StarTech.com Port USB Ethernet Adapter USB 3.0 USB C to RJ45
