RealSense technology

[Spyridoula_N_]

Hello,

I would like to know if it is possible to perform real time monitoring of the seabed profile. For example monitor the soil movement underwater, with a RealSense camera in a protective casing and get measurements of the depth of the seabed during the time of the experiment. The seabed profile changes as waves and current take action, resulting in depth variations. Would it be possible to get depth measurements of this constantly changing seabed profile? If yes how is this done? 

Thank you in advance.

Spyridoula

[Johannes_V_]

This is quite a special task you want to achieve. I dont think anyone here has any test data of a depth camera underwater. The R200 is pretty low price (100$). You should just try out. I am not sure if the range increases oder decreases underwater. However, I dont think you can scan anything reliable in a distance more than 8 - 10 meters.

Edit: maybe this helps https://www.youtube.com/watch?v=FZ6yE4HASls

[Spyridoula_N_]

Thank you Johannes V. for your response. I have seen the video you suggested, however in that case they are moving the camera around while I want it to be steady and monitor the moving seabed,  I think that it will be possible. Regarding the distance It won't be larger than a few meters (at most 2m). The range probably decreases due to the high attenuation of IR in water, however I am not familiar with the physics in this procedure. 

[MartyG]

This project reminds me of a mechanism I was designing to keep a smartphone and battery dry when submerged in the ocean.  The purpose of it was to measure the PH level in ocean water (it was for an ocean XPrize contest that my company didn't enter in the end).  It was based on the idea of a diving bathysphere.  I would put the electronics inside a plastic hamster ball that was pressurized with air, and the air bubble would keep the water out of the ball and maintain a dry, oxygenated atmosphere inside.

Anyway, I agree with Johannes that an R200 camera would probably be suited to this application.  It has a range of up to 5m in ideal conditions, I believe, and is "world facing" (it looks outwards towards the world, like the camera on the back of a tablet).

[Spyridoula_N_]

Thank you for your input Marty G. 

[jb455]

No idea if it'll work underwater, but yes it is possible (on dry land at least) to get depth measurements in real time. The depth image can be quite noisy even in ideal conditions though so you'd probably want to split the image up into different regions and get the average depth within those regions, rather than just looking at single points. It's fairly easy to do too.

You may need to use a brighter IR lamp to light up the scene if the camera's IR projector doesn't penetrate the water well, but if you do so you'll have to rely on stereo vision which doesn't work well on featureless areas (since the camera needs to match what it sees in both cameras so if there's no high-contrast features it can pick out it won't be able to join the images up to calculate the depth). So it's possible you'd need a more specialised bit of kit using SONAR or something, but you may have success with RealSense.

Good luck!

[Kilian_B_]

I don't believe it is possible to utilize the camera for the task you described. Infrared does not travel far underwater (the reason why the sea is blue after all)

[Gal_A_Intel]

Hi Spyridoula N,

The R200 RealSense Camera can only detect and scan up to 4 meters.
however we've never tried scan underwater.
If you are willing to try it, you can use Intel RealSense SDK samples in order to do your scan.

[Spyridoula_N_]

Dear all,

thank you for your responses. After trying the R200 underwater I can tell you that because of its range (0.4m) it can not sense anything underwater, however the SR300 can (I have not tried it but the authors of the paper "Underwater 3D Capture using a Low-Cost Commercial Depth Camera" told me I should use that one. I am mentioning this one here in case someone faces the same issue. 

Thank you all.