New Processing Options Let You Choose Speed vs. Quality
4–10x faster processing for large data sets, 2D orthomosaic maps
By Jeremy Eastwood, Mapping Technical Lead @DroneDeploy
No matter whether you’re building a house or planting a field, it’s important to use the right tool for the job. In the world of drone mapping, the “job” can vary dramatically. The subject matter that DroneDeploy users map ranges from small statues to tracts of land stretching over thousands of acres — and when it comes to processing that imagery, one size doesn’t fit all.
Today, we released new processing options to enable you to choose the “right tool” for your mapping job — giving you the option to dramatically reduce the time it takes to process your map.
Many of our users have told us that in some cases they want a map fast, and they don’t care as much about the 3D model or level of detail. Now you can choose whether to optimize your map processing for speed — ideal for large data sets and terrain — or for 3D model quality and measurements.
How it works
When you upload your imagery to DroneDeploy, just choose the “Speed”, “Quality” or middle option on the “Processing Options” slider.
Once your images are uploaded, the DroneDeploy Map Engine creates a three dimensional point cloud based on your images. From there, we use that point cloud to create a 3D model mesh, and generate a two-dimensional orthomosaic map by “flattening” the 3D model.
The processing option you select will determine how dense of a point cloud is generated based on your imagery, which will in turn affect the level of detail in your 3D model and elevation data. Select “Quality”, and Map Engine will generate a very dense point cloud and high-fidelity map and 3D model. Select “Speed”, and Map Engine will use a relatively sparse point cloud to generate your map and 3D model, significantly reducing the time spent on one of the most time-consuming parts of the photogrammetry process.
4–10x Faster Map Processing
The speed improvements we’ve seen in testing are dramatic. Especially for very large data sets, using faster processing can mean saving an entire day. This nearly 2,000 acre map in New Zealand took 24 hours to process using our standard processing, but when reprocessed using our fastest option, it took only 9 hours.
The exact processing time reduction will vary depending on a number of factors related to the size of your data set, your DroneDeploy plan, and whether you process imagery of terrain or a structure. In testing, we found that each increase in speed typically cut the processing time for a map by one half to one third. Below, see the time it took to process over 400 images into a map of oil storage tanks with each processing option.
Side-by-side Quality Comparison
When it comes to map processing, there’s a trade-off between speed and quality. Choosing a faster processing option will result in less detailed, or “lower quality” 3D data. To show you the difference in quality, let’s take a look at the oil storage tank maps.
Comparing the orthomosaic maps side by side, the differences aren’t easily discernible at first glance, but if you look closer you’ll notice some distortion around the edges of structures, particularly on the fastest map. When you map a building, it’s those crisp edges and fine details of the 3D model, that’s what you’re going to start losing and reducing when you move to faster processing. Overall, though, the faster processing is a great option for 2D mapping.
By contrast, the differences in quality are much more apparent in elevation maps and 3D models. In this example, there isn’t much change between standard and faster processing — just a little bit less crispness around the complex structures in the upper left corner — but data produced by the fastest processing option is significantly less detailed.
One thing to note is that the faster processing options reduce the fine-grained details of the map, but not the overall accuracy of the map. The points in the sparse point cloud are just as accurate as those in the denser point cloud, there are simply fewer of them.
Which Option Should You Use?
These options give you the control to choose what processing speed is best for your project, and we suggest that you experiment by uploading data sets with different settings.
In general, fastest processing will work best for projects like crop scouting and large-scale land mapping where there is a large data set to process, the map is needed quickly and the 3D and elevation data is less important. For 3D modeling of structure or for making accurate volume measurements, you’ll want to stick to our standard, high-quality processing.
Give it a try, and let us know how it goes ! Have feedback? Let us know on the forum.
Where to Learn More
To learn more about processing imagery into maps and 3D models with DroneDeploy, be sure to check out our support documentation.
Start Mapping with DroneDeploy
Want to learn how DroneDeploy can help your business? Visit www.dronedeploy.com to start your free trial or request a consultation with one of our team members. The DroneDeploy mobile application is available for free download for both iOS and Android devices.