The new sensor has two 100MP cameras. They are placed next to each other with a slight rotation.
It’s porpuse is to generate a extra wide covering area of photos during LiDAR captures.
The thing is that it is lacking the roll rotation, so instead of getting two cameras in the same position, one aiming right and the other left, we end up with both looking straigth down.
This messes up with the calibration process, specially in dense forest areas.
Only software that dealt with this so far is DJI Terra.
Thank you for your patience while we investigated this issue.
We have confirmed that the calibration difficulty is occurring because PIX4Dmatic does not yet automatically account for the fixed tilt angle of the L3’s dual-lens system. Currently, the software reads the gimbal orientation from the images but does not apply the necessary geometric offset for the specific angle of the cameras.
We have logged a development task to add full support for the DJI Zenmuse L3 dual cameras to resolve this.
Thank you for bringing this to our attention. Please don’t hesitate to let me know if you have any further questions.
I noticed the same issue. I tried to process data from the Zenmuse L3 and it worked great with DJI Terra but then I tried it with Pix4d Matic and noticed that the software does not account for the offset angles of the dual camera rig. Any idea when Pix4d may be able to supply a fix for this? Do you need any sample data?
Thank you for sharing your feedback. Please accept my apologies for the delay in our response.
I have confirmed with our R&D team that the request to support the offset angles for the L3 dual camera rig has been recorded. It is currently difficult to provide a specific timeline for when this fix will be implemented, but the team is aware of the requirement.
Regarding your offer for data, we already have the necessary datasets for testing, so there is no need to send samples at this time.
We appreciate your patience while we work on this improvement.
Has the issue been resolved? I saw that the new update to PIX4Dmapper already supports photos from the L3. Will the correct photo orientation and location work after updating PIX4Dmapper to the latest version?
Thank you for following up on the DJI Zenmuse L3 integration.
No, the current update to PIX4Dmapper does not resolve the photo orientation issues. While the recent update introduces initial support for the camera sensor, it does not support the offset angles. Please note that at the moment we have not yet planned to bring full L3 support to PIX4Dmapper.
The good news is that our development team is actively working on adding the proper, full support for the L3 sensor within PIX4Dmatic. This upcoming update will ensure the correct initial orientation for each input image. We expect this fix to be included in one of the next PIX4Dmatic releases.
Great to hear Pix4D is going to be supporting this sensor. After facing the same issue, we found that the roll offset was not included in the Exif metadata. Seemingly DJI Terra adds an offset from the gimbal for each LV and RV camera.
As a workaround, we used ExifTools to extract the gimbal angle and apply a +/-22.5° offset and either export a text file, or input back into the image Exif metadata. If anyone wants further information, please feel free to reach out at jason@geodronesurvey.com
Regarding the calibration (Nadir vs. Oblique), it depends on your flight pattern. I’ve included a link to our support documentation for the calibration templates here: Calibration - PIX4Dmatic
Concerning the 20–30 cm elevation offset from the LiDAR output: do you have any ground control points or checkpoints measured on-site? Checking this would help us confirm which data is actually correct and where the shift is coming from.
I’m using a standard map flight pattern, usually the Nadir would be the recommended settings. But since the L3 camera has an oblique 22.5 degrees angle in both cameras, I have doubts on what is the best option.
For mesh reconstruction the oblique got me a better result.
On the 20-30 cm elevation offset, the LiDAR is the correct one. We used checkpoints to validate it.
When runing the same dataset in DJI Terra, but as a photogrametry project, the generated point cloud matched the LiDAR. So my guess it’s something on PixMatic end and related to the L3 oblique cameras.
As for now, I’m using GCP and some marks from the LiDAR pc to get a good alignment.
Treating the dataset as Oblique makes sense due to the 22.5-degree camera angles, even on a standard map flight pattern. What are the exact calibration template and pipeline that gave you the best result with this sensor in PIX4Dmatic?
Concerning the 20–30 cm elevation offset, this does look like something we need to investigate on the PIX4Dmatic side. We definitely want to eliminate this extra manual alignment work from your workflow.
To investigate this properly, could you please open a support ticket and share this specific dataset with us?
These cookies are necessary for the website to function and cannot be switched off in our systems.
They are usually only set in response to actions made by you which amount to a request for services, such as setting your privacy preferences,
logging in, or filling in forms. These cookies do not store any personally identifiable information.
These cookies allow us to count visits and traffic sources so we can measure and improve the performance of our site.
They help us to know which pages are the most and least popular and see how visitors move around the site.
All information these cookies collect is aggregated and therefore anonymous.
If you do not allow these cookies we will not know when you have visited our site, and will not be able to monitor its performance.
These cookies may be set through our site by our advertising partner (Google).
They may be used by Google to build a profile of your interests and show you relevant adverts on other sites.
They do not directly store personal information but are based on uniquely identifying your browser and internet device.
If you do not allow these cookies, you will experience less targeted advertising.
