3

I work as a research assistant in the field of water resources management. We build small scale physical models of rivers and make experiments with sediment. The model rivers are about 20m long and 3m wide, the water depth is from 0 to half a meter. The bottom is either concrete or gravel.

One important task is to quantitatively measure the bedform that results after a model run. (i.e. where the gravel got washed away and where it accumulated)

I was wondering whether there is a chance to use a medical ultrasound device to record the topography of the (river) bottom. I'm hoping that these devices deliver a 2D profile of the riverbed and if I track to position + orientation of the ultrasound probe I can put together a 3d topography.

Has anyone ever done some experiments like that? I mean alienated a ultrasound device? or can anyone tell me why my idea is theoretically bad / good? Where could be problems? Is it possible to get spatial information out of an ultrasound image? I mean in the image the ground should become apparent very clear as there is only water between the sensor and the ground. I'm aiming at an accuracy of the resulting topography of about 5mm.

Mathias
  • 51
  • 2
  • 1
    If you're looking for the surface composition an ultrasound isn't really what you're looking for, look more for lidar and depth imaging cameras. – Dopeybob435 Sep 15 '16 at 12:54
  • the composition is not what we want - just the "first echo" or surface of the riverbed. Lidar or photogrammetry wont work in turbid water - unfortunately the water is often turbid... – Mathias Sep 15 '16 at 14:27
  • 1
    I'd think you'd be looking for some kind of sonar like a fish finder rather than a medical device. Medical devices are expensive, in large part because they're regulated so tightly. – DLS3141 Sep 15 '16 at 16:51
  • Fish finders also function in turbid water – DLS3141 Sep 15 '16 at 20:16
  • Thank you for yor answers! Fishfinders like from Lowrance or Humminbird have the downside that they only work in water depths of more than 0.5m. I've already tried a Lowrance device. As our model rivers are mostly very shallow ( few centimeters to maximum half a meter) these devices are ruled out. Also the resolution is not sufficiant for our needs. I really want to try a medical ultrasonic system - but wanted to ask if anyone can tell me if I maybe overlook sth and this wont work at all and I only waste time... – Mathias Sep 16 '16 at 07:23
  • You will likely need to modify/adapt a fishfinder to work in shallower depths, but it seems a closer starting point to what you're looking for. Again, any system engineered and approved for medical use will be at least an order of magnitude more expensive than an equivalent non-medical system – DLS3141 Sep 16 '16 at 13:33
  • What modifications would you suggest? I think modifying a fishfinder cant work out because they operate on much lower frequencies (khz) than medical ultrasound devices (mhz) do. (less relolution and importantly the "dead time" (time until a pulse leaves the transducer) is so high that it is not possible to record echos closer than about half a meter). Its a research project, so we have budget to buy a medical device - I'm just asking here because I want to avoid wasting money in case someone tells me why this is a bad idea from a technical point of view. – Mathias Sep 18 '16 at 13:13
  • have you reviewed the literature for related experiments? – agentp Sep 20 '16 at 00:41
  • Yes - unfortunately I found nothing... – Mathias Sep 21 '16 at 05:28

1 Answers1

2

I've had the possibility to test several different diagnostic ultrasound devices in a big bucket of water with gravel at the bottom! The result is interesting: At first, for all devices you have to use a "curved sound head", the linear one doesn't work at distances further than a few centimeters. With cheap devices I could hardly see the bottom or the gravel. With more expensive 3D / 4D devices the bottom and gravel became very clear and with the software we could measure distances with mm accuracy. The problem which however remains, is that the data cannot easily be exported for further processing. And there are to few parameters so adjust (not generally!, there are a lot... but for this special purpose). For example a tested "4D device" records a 3D model of the object when you scan it from different angles (e. g. for getting images of babys in 3D), but the time over which the images are taken is fixed and so the area of the riverbed which could be recorded is limited.

So basically the principle works - you can record the riverbed with mm accuracy but the devices are to much like "black boxes".

Mathias
  • 51
  • 2