Hello,

I come with a newbie question.
I want to enter a flight route on a M350 RTK and for it to follow it within 10 cm.
The base station is an EMLID. I guess it will need to be connected to NTRIP.

The drone is to fly approximatively 1.5 meters above water. It is pulling a stick with an underwater go pro, recording images for photogrammetry.

Is that doable?

Thanks for your help
Alex

No need additional NTRIP
10cm is a bit optimistic but not impossible

I recommend to disable anti collision feature

LV_Forestry Posted at 4-8 04:17
No need additional NTRIP
10cm is a bit optimistic but not impossible

Thanks for your reply.

What is realistic to expect ?
Is it possible to upload a manually generated flying route?

We do photogrammetry in shallow water and need to have the camera 1 -1.5 meters below the surface.
For testing the setup, we had a raft pushed by a swimmer with EMLID DGPS. But it takes ages to lawn mow a small area.
https://terra-submersa.github.io/coordinates-lawnmower/

alex mass Posted at 4-8 04:38
Thanks for your reply.

What is realistic to expect ?

It all depends on the wind and the trajectory.  I'm talking about the precision of following a pre-established flight path.  Because yes its possible to draw a line, on QGIS, Google Earth... and import it in the drone.

If what interests you are the coordinates of the drone itself then you can hope for accuracy of the order of centimeters.  

Be careful if you synchronize the GPS coordinates of the drone with those of the GoPro, I don't know what influence 1.5m of water has on a GNSS receiver.  Probably not ideal.  

And I have a question, what about the image distortion caused by moving the camera in the water?  Is there ?

very interesting project.  Have you ever tried echo sounders?

Thanks again.

The positioning of the camera is another problem in itself. Getting the correct timestamp on the picture (https://github.com/terra-submersa/coordinates-label-photos/tree/main/experiments/do-we-have-the-correct-position). We will have to do some post-processing based on the recorded angle of the stick.
And to remove unnecessary images: https://docs.google.com/document/d/e/2PACX-1vSS9lody_pzMpdDJfpkczp9-Z65DgvwbtN69H3uTy3hhaqFMJcmw3We1s6h_DMtRjbe-ZY3d6Ii8xiF/pub

Any comments are more than welcome
Alex

And echo sounder would not work for some site, where the detected wall are either very subtle, or because we need image to find out what is an original artefcat or something  else.
Moreover, on some sites, the walls appear only a few centimeters over the sand surface.

alex mass Posted at 4-8 05:28
And echo sounder would not work for some site, where the detected wall are either very subtle, or because we need image to find out what is an original artefcat or something  else.
Moreover, on some sites, the walls appear only a few centimeters over the sand surface.

Then bathymetric lidar. But its not the same range of price as GoPro.

Yellowscan has released a fantastic device at beginning of this year.

impressive !
And there are right down the corner...

Before UAS, I spent 20 years doing nearshore bathymetric mapping using various sonars, ROV, etc. The new Yellowscan bathy lidar is definitely a breakthrough, being the first such payload that can be carried by sUAS.

I've done a bit of U/W photogrammetry as well. I'd be surprised if you can achieve what it is you're after using UAS. I understand the desire to speed up the data acquisition over what a swimmer can achieve, but I don't think you're going to be able to go much faster using a drone and still get the required number of images with the needed overlap and image quality. Seems like a pretty significant risk of a not inexpensive drone for what might be a minimal efficiency increase. But good luck if you try it! Personally I'd look at a small skiff with an electric trolling motor or somesuch. Cheaper, safer, and simpler. If you want to get "fancy" and keep folks out/off of the water you could fashion or buy of the shelf a small autonomous boat (ASV).

https://www.seafloorsystems.com/echoboat-160
https://www.seafloorsystems.com/echoboat-240

(I'm a customer only, and have no financial interest in the product)

Regarding GNSS coordinates, you don't even need them for the images to do photogrammetry, but it is nice to have. GNSS satellite signals do not penetrate water so you'll be calculating image position from that of the drone not the GoPro, as you mentioned.

patiam Posted at 4-8 07:45
Before UAS, I spent 20 years doing nearshore bathymetric mapping using various sonars, ROV, etc. The new Yellowscan bathy lidar is definitely a breakthrough, being the first such payload that can be carried by sUAS.

I've done a bit of U/W photogrammetry as well. I'd be surprised if you can achieve what it is you're after using UAS. I understand the desire to speed up the data acquisition over what a swimmer can achieve, but I don't think you're going to be able to go much faster using a drone and still get the required number of images with the needed overlap and image quality. Seems like a pretty significant risk of a not inexpensive drone for what might be a minimal efficiency increase. But good luck if you try it! Personally I'd look at a small skiff with an electric trolling motor or somesuch. Cheaper, safer, and simpler. If you want to get "fancy" and keep folks out/off of the water you could fashion or buy of the shelf a small autonomous boat (ASV).

Thanks for your comments and pointers.
You are perfectly right, the flying speed will not be much faster than a swimmer. We spent many (and many !) hours in the water, pushing a raft with a stick to acquire photos. And swimming within a 30 cm lane for hours is not an easy task.
Positioning the camera is described in the links above. Having a precise location is key for the reconstruction computing times (the area we aim at is 100x100 meters).
Finally, water visibility changes pretty fast in our working area. When conditions are met, we aim to survey the largest possible area the soonest.
Some tried the boat, but it offers other constraints.
And we will give a look at Yellowscan! Even if we don't look at the price tag for the moment (it certainty is heavy). For some archeological sites, it might really be a killer. While we still need to challenge the image precision for others.
Thanks again,
Alex

Glad to offer ideas.
I have used the EchoBoat 240 USV that I linked and it is very maneuverable, easy to pilot manually, and can be programmed for autonomous missions the same way as UAS. It has all day endurance, RTK GNSS/INS and a moon pool that can hold a sonar payload or could be used to house your camera on a retractable pole. Bathy resolution with a good shallow water MBES would be in the cm range, not quite as high as with photogrammetry but not dependent on light or good visibility. You could build your own, rent or buy. There are of course other makes and models but this is the one with which I have recent experience.



Just an alternative to UAS. I understand that generally the tool you have is better than the one you don't, but I'd hate to see a good UAS take an unplanned swim as they can usually only do that once

Hi, is it possible to get RINEX data from a Matrice 350RTK when I do not use a DJI payload? Or is that only possible, when I am using L2 or other DJI payloads?

I'm not sure of the answer for this, as I do not own an M350. But in theory, the aircraft's ability to create the files needed for PPK should not be at all dependent on the payload. The GNSS system and its functionality should be payload agnostic. It should work with no payload at all.

HOWEVER, see this thread:

Matrice 210 RTK V2 - No PPKRAW or Rinex.obs data recorded
It appears that at least for the M210RTK, "should" is not enough. DJI seems to have made at least that platform unable to record data for PPK unless certain payloads are used. Works with X7, but not X5S, for example.

This is a terribly flawed design.

As I said, I don't know if this applies to the M350 as well. Hopefully someone with direct experience will chime in. But I wanted to at least share some information regarding a predecessor of the M350.

Gernot Weyss Posted at 4-10 03:41
Hi, is it possible to get RINEX data from a Matrice 350RTK when I do not use a DJI payload? Or is that only possible, when I am using L2 or other DJI payloads?

I confirm, unless having a payload that calls the function to generate .OBS file, I have never managed to find such a file in the memory of the drone itself.

For third party payloads I use a U-blox F9P receiver which records the obs in a file.

Just another example of how "DJI Enterprise" is an oxymoron...