Drone orthomosaics are great … but what can I do with them?
In the latest Soar Cast, Drone Orthomosaics — What are they? Insights from an expert, Darren Smith interviewed Chris Strasbaugh, an Educational Technologist and UAV Coordinator at Ohio State University College of Engineering. They discussed how to create drone orthomosaics, what is photogrammetry, best practices of using the SOAR+ map hosting and much more.
Chris Strasbaugh has been working at the Ohio State University College of Engineering for four years, where he launched a successful drone program at the Knowlton School of Architecture. His drone program was used in landscape architecture, as well as city and regional planning courses. Over time, he obtained a drone pilot’s license and flew drones to generate 3D models, study poverty disparity and transportation networks. In only four years, the program took off and recently was expanded.
While photogrammetry as an academic field is well-known within a broader scientific community, Strasbaugh is often asked what a photogrammetrist does. In such cases, Strasbaugh answers that it’s the process of taking multiple photographs with a strategic overlap and merging them with specialized software into one large photograph. Because you end up with a georeferenced photo, you can use it for doing spatial measurements or add spatial data on top of it.
Strasbaugh is very enthusiastic about the possibilities of georeferenced imagery for students. He explains that the possibilities are endless:
Strasbaugh: “Specialized drone image processing software can also generate an accurate 3D model of an area, which allows for doing measurements in 3D. Those can be combined with spatial datasets such as thermal or multispectral data. Practically, the sky is the limit. It’s just amazing what you can do with a series of drone photos that are stitched together.”
One thing that Strasbaugh has noticed doing many drone imagery projects, is the growth of datasets over time. Using GeoTIFFs as a standard data format for creating orthomosaics, multispectral imagery, and surface models, a project covering 10-acres would result in one gigabyte of data that could be managed very well on a laptop. However, data sizes quickly increased up to as many as eight gigabytes per GeoTIFF file and got even larger when different projects were merged, which became a major drain on the IT system.
The ideal solution would be able to share large data files inside and outside the organization, to provide access to different research and community groups. Strasbaugh tried multiple image platforms in the cloud, among them being the Soar+ map image and raster image sharing platform, launched late 2019. Interestingly, he used it differently than its creators had in mind: rather than being a cloud data storage solution, he thinks it can be used as an online platform where different data layers are stacked on top of each other so that different maps with different types of data can be created over time.
While he tried several cloud storage platforms, the Soar+ platform stands out for Strasbaugh for multiple reasons. He and host Darren Smith also share some ideas of functionality to include at a later point in time into the platform:
Strasbaugh: “Soar+ is the only place where you can stack things well. Being able to visualize and select these layers has been super useful. Also, it’s more open than other solutions in the market.
Smith: Thanks Chris! I’d be very keen to see multiple datasets in the platform. For example, we’re using RGB cameras and there are several data formats that you can from that in terms of stitched imagery or orthomosaic or image in a terrain model. This allows you do create a ‘greeness’ vegetation index (VARI) to derive the relative plant health. *VARI is a alogorithm using the red/green/blue components of visual light; VARI=(Green-Red)/(Green+Red-Blue).
However, there’s a data format problem that needs to be solved first. When first accessing, measuring and zooming in and out of GeoTIFF files in the cloud, Strasbaugh discovered that GeoTIFFs were unwieldy. Next, he started looking for a different format that could overcome these, so that data can be exchanged no matter their size. One good candidate is the ECW format, however, this is a proprietary format that requires off-the-shelf software to convert it to that format.
At the moment, Strasbaugh is looking into JPEG2, an open data format that is comparable to ECW:
Strasbaugh: “Just like ECW, JPEG2 is a compressed data format but has georeferencing built-in to it. While the original JPEG data format threw out data for compression purposes, which means you can never get it back. JPEG2 was different, in the sense that it was specifically designed as a lossless compressed data format, meaning you’re not losing data but only temporarily hiding it until you need it again. Although JPEG2 hasn’t seen wide adoption, that doesn’t mean it’s any less successful. I think it has some potentials in the photogrammetry space because the compression is quite sizeable without losing the data. Also, it’s accessible in Photoshop. There are some things we’re working out at the moment before making the transition.”
