I just completed an aerial survey and attempted to process the images but have received an error: Exception thrown: Too many missing images leading to incomplete captures . There seems to be lots of images to cover the area. Any thoughts about how I get past this error? Thanks.
Hey Paul,
I would identify all those images then remove those images and re-create your project from the beginning without those images and it should work. I have seen the same issue in another post, I believe you could also use the earlier version of Pix4D version 4.2.17 as an alternative.
Hello,
Was there ever a solution found for this?
Would using version 4.2.17 solve the problem even if the project was created in the newest version ?
Hi,
When Sequoia camera is triggered, a capture per each sensor it is taken. These captures are assigned to one rig instance to calibrate the rig successfully. If some images are missing from a capture this will be discarded. In case more than 5% of captures is incomplete then the software does not further proceed with the calibration.
However, sometimes it may happen that no images are missing but they share the same unique id (8/12 instead of 4 each rig). Unfortunately, the issue is related to Sequoia’s firmware. The only solution for such datasets is to process it with the previous 4.2.17 that is using the capture time to assign the camera to a rig instance
In the versions, 4.2.25 and later, the images are assigned to the capture based on capture id. The reason it was changed is because this allows consistency in the processing in case images are missing (but less than 5% captures). But if wrong tags are written by the camera and 4 images instead of 8 are assigned to a capture, then you need to process the images based on capture time (4.2.17).
Hi Momtanu,
These images weren’t captured with a sequoia rather the Red Edge. Will this method of downloading the previous version solve the problem?
Hi,
The rig assignment algorithm has been changed for both sequoia and rededge. As I explained before, the new algorithm uses the unique identifier associated with each capture, available from the image EXIF of the images, to group them and assign it to rig instances.
Only a limited amount of incomplete captures is allowed for the processing to continue. This limit is 5% of the total number of captures.
Could you please check if the images from for which you had received the message are complete and let me know? You can then delete them and then proceed.
If there is no image duplication (in the naming and rig assignment) and it is a case of missing images, then working with 4.2.17 might not work. In that case, if you use 4.2.17, since it is using capture time for assignment, it will mess up the project. For example, if you have images (two captures, capture1 and capture2): red1, rededge1, NIR1, blue1, red2, rededge2, NIR2, blue2, green2 (look green of capture1 is missing). The algorithm of 4.2.17 will assign green2 to capture1. But if it is an issue with duplicate naming, 4.2.17 will work as it will assign the images to the capture based on time. Hope I am able to make this clear.
Hi,
I’m having the same issue with a dataset captured with the DJI P4 Multispectral.
I’ve checked and all the images is there, no images missing.
Is there a way to solve this? New image acquisition. is impossible.
Hello @Biodrone, Can you forward the .p4d and the log file of the project? I would also suggest you to use the different calibration methods and see if that is helpful.
Hi Kapil,
Thank you for assisting.
Here’s the .p4d file.
We’ll try some of the other methods to check if that could be a solution.Helgoya 2.p4d (1.3 MB)
Hello @Biodrone, I would need a log file to know about the issue? Did you try the different calibration methods(i.e. Standard Calibration method) to see if that was helpful?
Yes, here it is.
All calibration methods are tried, none worked.
The program seems to think images are missing because the images are not sorted in the correct order. Usually the RGB- images are displayed first, then the other bands after that. When importing the images from the drone, I noticed that the images looked shuffled. Could that be an explanation for why the program thinks images are missingMultispektral test.log (1.3 MB)
Multispektral test.p4d (1.3 MB)
?
Hello,
The log file shows that only 2 or 3 images were captured instead of 5 bands images per image capture, which is the issue. Can you confirm that you have five bands per capture? If so, can you forward me the dataset so that I can look into it and see if there is anything we could do from our side?
