Please select Into the mobile phone version | Continue to access the computer ver.
Issue with 2D reconstruction on DJI Terra 3.9.0
1114 9 2023-10-26
Uploading and Loding Picture ...(0/1)
o(^-^)o
Robert Marshall
lvl.2
Flight distance : 745154 ft

United Kingdom
Offline

Hi Wanda,

I'm seeing an odd artifact on my 2D renders, colour tone banding.

It is hard not to see and i've got a customer refusing to accept the work due to this.

I'm using 3.9.0.

The banding takes the same direction as the flight of the drone.

The source images seem to be okay other than some cloud \ non-cloud shots here and there.

The 3D model does not show these artifacts.

It is the same M300 and P1 that fed previous large-scale renders that didn't show up this problem on down-level versions of Terra.


Is this a bug?

BandingEffect2.png
BandingEffect1.png
2023-10-26
Use props
LV_Forestry
First Officer
Flight distance : 4726654 ft
Latvia
Offline

Its a well known issue with Terra.
Try with serious photogrammetry software, the artifact should disapear.
You can send me your dataset, then i could try in Metashape to let you appreciate the results. Therefore you can assess if the trouble doesn't come from camera settings.
2023-10-27
Use props
Robert Marshall
lvl.2
Flight distance : 745154 ft

United Kingdom
Offline

LV_Forestry Posted at 10-27 00:26
Its a well known issue with Terra.
Try with serious photogrammetry software, the artifact should disapear.
You can send me your dataset, then i could try in Metashape to let you appreciate the results. Therefore you can assess if the trouble doesn't come from camera settings.

Interesting offer, let's connect ...
2023-10-27
Use props
LV_Forestry
First Officer
Flight distance : 4726654 ft
Latvia
Offline

Robert Marshall Posted at 10-27 00:51
Interesting offer, let's connect ...

I wrote down the email address.  I recommend that you remove it from the forum, unless you are ready to receive spam.  
You can prepare the dataset on a cloud.  I'll contact you when I'm in the office.  Probably tonight.
2023-10-27
Use props
Robert Marshall
lvl.2
Flight distance : 745154 ft

United Kingdom
Offline

LV_Forestry Posted at 10-27 00:57
I wrote down the email address.  I recommend that you remove it from the forum, unless you are ready to receive spam.  
You can prepare the dataset on a cloud.  I'll contact you when I'm in the office.  Probably tonight.

I could instead kick off a trial of MetaShape to discount the camera, and put the same imagery through it.

I think that might be better rather than this becoming a time-sink for you.

Ping me an email so we can still connect, and I'll let you know how it goes.
2023-10-27
Use props
LV_Forestry
First Officer
Flight distance : 4726654 ft
Latvia
Offline

Robert Marshall Posted at 10-27 01:03
I could instead kick off a trial of MetaShape to discount the camera, and put the same imagery through it.

I think that might be better rather than this becoming a time-sink for you.

You can try with the trial version too. Results will be basically the same if we use same settings. I simply need to put pictures in the software and run my script. Tomorrow when i will get up the raster will be ready.
2023-10-27
Use props
Robert Marshall
lvl.2
Flight distance : 745154 ft

United Kingdom
Offline

LV_Forestry Posted at 10-27 01:08
You can try with the trial version too. Results will be basically the same if we use same settings. I simply need to put pictures in the software and run my script. Tomorrow when i will get up the raster will be ready.

Cool, I'm doing the trial right now.

I will wait for you to email me. I'm in UK time zone.
2023-10-27
Use props
DJI Susan
Administrator
Offline

Hi there, sorry for the issue. For your case, as the water surface is not diffusely reflective, the water surface of the same area will differ greatly in the same positions of the aircraft. In addition, the parameters shot by the same camera will also differ, resulting in a large difference in brightness in the shot images. As such, obvious traces of splicing may occur. In addition, crops taking up a large area will also generate a certain non-diffuse reflection. Obvious traces of splicing may also be shown in the 2D reconstruction of field scenarios.

You may try the following suggestions:
1. While shooting, set the metering mode to global metering. This can solve the issue to some extent.
2. Avoid image data collection when the sunlight is strong (such as at noon).
3. The algorithm will be optimized in later updates.
4. Enable Light Uniformity to alleviate the issue to some extent.

Hope this helps.
2023-11-11
Use props
Robert Marshall
lvl.2
Flight distance : 745154 ft

United Kingdom
Offline

DJI Susan Posted at 11-11 00:22
Hi there, sorry for the issue. For your case, as the water surface is not diffusely reflective, the water surface of the same area will differ greatly in the same positions of the aircraft. In addition, the parameters shot by the same camera will also differ, resulting in a large difference in brightness in the shot images. As such, obvious traces of splicing may occur. In addition, crops taking up a large area will also generate a certain non-diffuse reflection. Obvious traces of splicing may also be shown in the 2D reconstruction of field scenarios.

You may try the following suggestions:

I'll take that onboard and try some of those suggestions thank you Susan.
2023-12-8
Use props
DJI Susan
Administrator
Offline

Robert Marshall Posted at 12-8 12:33
I'll take that onboard and try some of those suggestions thank you Susan.

You are welcome, hope it helps with your situation.
2023-12-11
Use props
Advanced
You need to log in before you can reply Login | Register now

Credit Rules