Matrice 4E digital zoom stuck at 1.01 or 1.12

Dave01 · Friday at 2:50 PM

Good afternoon everyone,

To introduce myself a little, I am Dave, I use drones to study the spatial ecology of marine species. I've used the phantom and mavic in the past to do some standard mapping work to make 3D models and orthomosaics. Now, I've recently got a hold of the Matrice 4E to undertake a new project to measure some small critters in the intertidal - I'm new to this drone, and also to using RTK so I'm on a learning curve at the moment. I have to fly close to cliffs and use the 70mm or 168mm camera using a manual flight (imagine you have to move the drone up and down with the rhythm of the waves - quite high risk surveys).

I'm having a major issue at the moment, I am using these images to create 3D models in Agisoft. When I import the images into Agisoft, I get lots of warnings about digital zoom. which is especially problematic in general and was amplified when I tried to use images from the 70mm and 168mm in the same project, it's a mess!

After further investigation, the Exifs show that the 70mm has a digital zoom of 1.12 and the 168mm has a digital zoom of 1.01. No matter what I try, I cannot seem to set the zoom to 1, I have tried it manually with the wheel, I've tried using the R2 and R3 button to set the zoom level, neither seem to make a difference. The wide photos are locked nicely at 1 (ie no digital zoom).

Has anyone got any ideas how I can deal with this?

In an ideal world, I am trying to use the zoom cameras to make the models, but also supplemented with some wide camera images to make sure the whole scene clicks together in the case my zooms didn't get the sufficient overlap dropped (as mentioned its a fully 3D scene with manual flights)

Also a quick add on question, as mentioned I am new to using RTK - if I have RTK enabled and connected to the drone (through either NTRIP or with the RTK3 station), what are the coordinates in the image EXIFs? GNSS or RTK positioning?

Thank you for your time,
Kindest regards,
Dave

Dave01 · Friday at 2:59 PM

Perhaps I forgot to mention, on the screen, the zoom is set to x3 or x7 exactly.

jaja6009 · Friday at 3:27 PM

Here is a link to Agisoft forum. You can have different zoom levels/focal lengths in the same chunk. I always thought you could not.

Different camera, same chunk

www.agisoft.com

You will have to make sure your overlap is ok since the different zoom levels may make the automated flights parameters not provide the correct overlap.

I posted a response that I deleted since I was unaware that Agisoft now will sort images with a different focal length into their own groups.

On your RTK question.

The images will be in the same datum/coordinates system that the base sends them in.

If you use an RTK/NTRIP subscription service that uses NAD83(2011), then that is what your images will be tagged in.

Basically when using RTK or PPK, the drone receives and uses the same datum/system that the base sends.

If you use a DRTK3 then the DRTK3 will send whatever it is setup on. If you set it up on an NGS benchmark in NAD83(2011), then that will be what your images are in.
But, if you just setup the DRTK3 and average a point, this point will not be centimeter accurate to a datum/system. But if you always use this point then you will have relative accuracy, as in being relative accurate to this point.
You can however have the DRTK3 receive corrections from an RTK subscription service or as I said setup on an NGS monument and then have absolute accuracy in NAD83(2011).

The above assumes you are in the United States and need NAD83(2011).

The above also holds true for PPK.

Dave01 · Friday at 5:03 PM

I have an email saying there is a response to this thread, but I can't see any posts...

For further context, I'm attaching screenshots of the zoom issue...

jaja6009 · Friday at 8:40 PM

I replied, but after visiting Agisoft I was incorrect. Also my other reply has a link in it to Agisoft and is waiting for a moderator to approve it.
But i posted below my response on RTK.

On your alignment issue:

You can process the different cameras together in one chunk in Agisoft.

What happens after the first alignment?
If you are flying a mapping mission in Pilot 2 and then zoom in, your overlap will be off a lot. This could affect your initial alignment.

I have never combined different cameras in one chunk, I usually would make a separate chunk for each camera and then combine them, but you don't have to follow this workflow.

If you are not having any luck try to align each group of images in its own chunk to see if any group is giving problems. If a particular zoom level of images is not aligning then that is one of your problems.

On your RTK question.

The images will be in the same datum/coordinates system that the base sends them in.

If you use an RTK/NTRIP subscription service that uses NAD83(2011), then that is what your images will be tagged in.

Basically when using RTK or PPK, the drone receives and uses the same datum/system that the base sends.

If you use a DRTK3 then the DRTK3 will send whatever it is setup on. If you set it up on an NGS benchmark in NAD83(2011), then that will be what your images are in.
But, if you just setup the DRTK3 and average a point, this point will not be centimeter accurate to a datum/system. But if you always use this point then you will have relative accuracy, as in being relative accurate to this point.
You can however have the DRTK3 receive corrections from an RTK subscription service or as I said setup on an NGS monument and then have absolute accuracy in NAD83(2011).

The above assumes you are in the United States and need NAD83(2011).

The above also holds true for PPK.

Dave01 · 2025-12-01T14:25:06-0500

Hi jaja6009,

Thanks for your reply, appreciate it!

I was also unaware that Agisoft would group by calibration. In the past I found any (accidental) zoom wrecked havoc in the optimising camera step - I will have to have a play around with that. In any case, my last set of trials I decided to just stick to one zoom and just fly the drone a little further away from the subject as a failsafe to ensure the images can stitch.

So in the end, the Matrice constantly being stuck at 1.01, technically shouldn't be an issue? (my only concern is Agisoft trying to link a 1.01 with the location of the drone and getting the calibration off - the area is too extreme to place GCPs)

Regarding the RTK, great to know that position in the images are already updated to what it's being fed (NTRIP or DRTK3).

I'm still working out my best method using this, absolute positioning isn't as important, but relative position is. I have a RTK subscription but in some locations of my work, I am too far away and behind cliffs so I've been struggling to get signal either for the drone directly or the DRTK3. So at the moment, I'm using the average DRTK3 position but I need to improve my workflow a little bit to get this right - I may be back here in the near future with questions hehe, but for now I'm still very much in the experiment and tinker phase! I am based in Europe btw

Kindest regards, and thanks again,
Dave

jaja6009 · 2025-12-01T16:49:59-0500

Dave01 said:
Hi jaja6009,

Thanks for your reply, appreciate it!

I was also unaware that Agisoft would group by calibration. In the past I found any (accidental) zoom wrecked havoc in the optimising camera step - I will have to have a play around with that. In any case, my last set of trials I decided to just stick to one zoom and just fly the drone a little further away from the subject as a failsafe to ensure the images can stitch.

So in the end, the Matrice constantly being stuck at 1.01, technically shouldn't be an issue? (my only concern is Agisoft trying to link a 1.01 with the location of the drone and getting the calibration off - the area is too extreme to place GCPs)

Regarding the RTK, great to know that position in the images are already updated to what it's being fed (NTRIP or DRTK3).

I'm still working out my best method using this, absolute positioning isn't as important, but relative position is. I have a RTK subscription but in some locations of my work, I am too far away and behind cliffs so I've been struggling to get signal either for the drone directly or the DRTK3. So at the moment, I'm using the average DRTK3 position but I need to improve my workflow a little bit to get this right - I may be back here in the near future with questions hehe, but for now I'm still very much in the experiment and tinker phase! I am based in Europe btw

Kindest regards, and thanks again,
Dave

On Agisoft I think your biggest issue if using one of the zoom cameras will be not having the correct overlap. Try to increase overlap.
Also as you mentiuoned keep the zoom at the same exact level.

You can try PPK also. After your flight download the RINEX logs from your image folder and then from your Base (CORS, RTK subscription service) and then process in RTK lib, Redbox or Emlid.

One last question, how far away are you from your subject material? The wide camera still can give good GSD and resolution at 200 feet away.

And maybe try out one of the facade modes or even Smart Capture where if you can use RTK even with your own base next to you, you can literally fly within feet of the subject material. I do it on buildings, and while it is nerve racking, the RTK precise positioning will keep the drone where it needs to be.