Depending on the dataset you upload, it can take anywhere from an hour to several days to process the orthomosaic, DSM, and DTM. You will receive an email once these have been created.
If your dataset has not been processed in a reasonable time according to its size, then below are some of the common causes of extended processing times or failures in the processing. Feel free to reach out to us on our message bot if you need more specific information on your case.
The most common reasons for a dataset not processing are:
Not all the photos are uploaded.
Related to the content in your images, and how difficult it is for the computer algorithms to process them.
Let's explore these in more detail.
Dataset size
The larger the dataset, the longer it takes to process. Datasets smaller than 500 images typically take under three hours to process.
Those on a free subscription (Essentials) can upload a dataset of up to 500 photos.
Those on a paid Professional subscription can upload a dataset of up to 3000 photos.
If you have a dataset larger than 3,000 photos, you will need to contact us to subscribe to a Pro + subscription.
Not all images are uploaded
Our automated process of creating your orthomosaic starts once 100% of the photos have been uploaded. If 100% of the photos are not uploaded, then the processing workflow will not start.
When you upload your photos to GeoNadir, there are two stages that occur:
When you drag and drop or select the data from your computer to upload we read the thumbnails, metadata, and number of images you are uploading. This generates the details needed for the upload card.
When you push upload, we then start to transfer the data from your computer to our cloud server. Once 100% of the images have been uploaded as read in the first stage, then the processing of the orthomosaic will commence.
You can see the number of photos uploaded by going to the dataset card (under Datasets) and then looking at the 'number of images' under Technical details. If this number is lower that the number of images for this mapping mission on your computer, and the upload is no longer occurring, then there has been a failure in the upload.
You will have to delete this dataset and re-upload as a new dataset. There is no way currently to add more photos to a dataset once the upload has started or completed/failed.
Lost internet connection
If the internet connection is lost during the upload process, we will attempt to resume the upload process once the computer comes back online.
However, if the browser is refreshed, closed, or the computer turned off, then this upload process will fail. Note that some browsers will 'pause' tabs that are using a high amount of bandwidth and are not in the active viewing tab. Once the tab then becomes the 'active tab' again the tab is reloaded - this will cause a failure in the upload process.
It is best to have one tab uploading at a time, have this tab 'active' on your screen, and if you have to leave your computer for an extended version make sure the computer doesn't go to sleep by playing a YouTube video on repeat or putting a PowerPoint in presentation mode in the background.
Insufficient overlap and / or sidelap
As we process your data, the computer is looking for matching features (pixels) in areas of overlap and sidelap between sequential photos. If the overlap and sidelap is insufficient, the computer can't find enough matching features. We recommend using 80% overlap and 80% sidelap for all missions.
Large, homogenous areas (e.g. water)
If there aren't many distinctive features in your imagery, the computer finds it difficult to match similar pixels. Large areas of water, grasses, dry agricultural fields, or concrete sometimes have this challenge. Try flying at a higher altitude as this will often mean that each photo will encompass a greater variety of features that the computer can recognise.
Glint or glare on water bodies
In cases where there are large water bodies that are highly reflecting sunlight, the computer processing algorithms view these areas as homogenous and are unable to find matching features. Try capturing your data earlier or later in the day to avoid glint, particularly in the middle of your photos. Increasing altitude can also mean that each photo will encompass a greater variety of features that the computer can recognise.
Photos are oblique
Our processing algorithms are optimised for nadir (top-down) imagery. While we can process datasets with some oblique images, for best results make sure that you have covered your entire site with the drone camera pointing directly down.
Photos are overexposed or underexposed
If photos are too bright (overexposed) or too dark (underexposed) the computer will have difficulty finding matching features in the photos. Have a look at the camera white balance and ISO settings - usually automatic works well as a starting point.
Photos are out of focus
If features in the images are blurry, it will be difficult for the computer to identify matching pixels. Focus issues can arise from flying too fast, or having the incorrect focal point. Usually autofocus is the best setting to use, though try infinity focus if you are having difficulty with autofocus changing throughout a capture. If you use a mapping mission planning app (preferable), you shouldn't have an issue with speed, though it's worth checking this parameter as well. Finally, check that the lens is clean.
Feature within the images are moving
This is often an issue on windy days with trees or grasses that move around a lot (e.g. palm trees, sugar cane, wave action on the water). Sometimes this will result in duplicate features in an orthomosaic, but in extreme cases it may mean that the orthomosaic can't be created at all.
Want to learn more? Check out our range of training options to help you take your drone mapping skills to the next level!