Here’s a blog post for the drone professionals and enthusiasts out there, straight from the Finnish Geospatial Research Institute FGI’s leading UAS experts, explaining how you can benefit from using a test field for your drone cameras and sensors.
Drone based remote sensing is a rapidly increasing field of science and business
Drones are powerful tools for measuring and observing objects and different phenomenon when equipped with cameras, videos, thermal cameras and other sensors. Their application potential is practically unlimited.
A lot can be done by visually evaluating the collected data, such as camera and video. Even in this case the user has to find the best flight options using test flights.
Increasingly, drones are used for quantitative, automatic and autonomous analytics. In this case, the drone is a sensitive measurement device that has to be properly calibrated and validated.
On the accuracy of the remote sensing measurements
In a drone remote sensing application, the user has typically four major concerns on the data quality. We call them the four quality indicators:
- Can I trust that the measured objects are at correct locations and that the geometric measurements, such as volumes, distances and sizes are correct? This is referred as the geometric accuracy of the system.
- How small objects can I observe and analyze, in other words, what is the spatial resolution of my system?
- How about the spectral and radiometric accuracy? As in, are the colors and intensities in images representing real object properties and how small details can I observe?
- What is the accuracy in my application, for what I actually want to measure, be it estimating crop biomass, pile volume or tree height? We call this the end product accuracy.
Picture: An aerial view to the Sjökulla image quality test field with permanent radiometric and bar targets, and portable gray scale radiometric targets and Siemens star.
Picture: Close up to permanent and portable reflectance targets.
Three different types of test fields are used for our research. What are those?
At the FGI, we categorize the test fields to 1) permanent and 2) temporal remote sensing system test fields as well as 3) in situ (application) test fields. All types of test fields have targets that can be used for calibrating and validating remote sensing system for some or all of the four above mentioned quality indicators.
A well-known permanent test field is the Sjökulla test field in Kirkkonummi Finland. It has been established already in 1994 and companies and National Land Survey of Finland have used it for calibrating and validating airborne mapping systems. It has currently approximately 50 permanent ground control points, with geometric precision of 5 mm or 10 mm. It has permanent radiometric targets of size of 14 m by 14 m for radiometric calibrations and validation. For spatial quality evaluations, permanent line bars and edge targets are available.
A temporal remote sensing test field can be established at desired place when needed. Ground control points are measured using static or real-time kinematic (RTK) GNSS technique. For radiometric calibration, a portable gray scale with seven intensities of 5-70% reflectivity and size of 5 m by 5 m, or smaller portable 1 m by 1 m panels are used. For spatial resolution assessment, a portable Siemens star with a diameter of 13 m is available.
In situ test field is typically deployed during a single remote sensing campaign. Common practice is to use three or more ground control points measured by RTK GNSS and a minimum of two reflectance panels to calibrate image digital numbers to reflectance factors. This procedure is commonly used by researchers and companies in the practical drone remote sensing tasks.
Picture: In situ reflectance panels in shallow seabed mapping application study. Photo by Lauri Markelin.
Why are test fields needed?
Performance of a drone remote sensing system is dependent on many factors, such as the sensor and the drone used, the gimbal, the flight speed, atmospheric conditions, and the last but not least, the data processing software and its settings.
When aiming to use the system in quantitative measurements, it is compulsory to know the system performance. The remote sensing system test fields are where we get to know our tools.
Test fields have also an important role when developing next generation drone remote sensing systems. As an example, for practical applications, we are aiming to minimize the need for in situ targets, which require a lot of work to install. They are a bottleneck for autonomous use and beyond visual line of sight operation. Therefore, we are developing so-called direct drone remote sensing systems that do not require such in situ-targets. Currently we are evaluating the use of well calibrated camera systems together with precise RTK GNSS and FINPOS positioning service for direct geometric measurements. We are also developing precise techniques for direct reflectance measurements based on a radiometrically calibrated camera and simultaneous onboard irradiance observations.
So what is then the best way to use the different kinds of test fields? Our advice is to use the permanent or temporal remote sensing system test fields during the development and deployment phase to calibrate and validate the system as well as to check the system whenever necessary. The use of in situ test fields are recommended in typical application campaigns to confirm the reliability of the data. And as soon as reliable direct measurement systems are available, the need for in situ targets will be reduced.
We welcome you to test your systems in the FGI’s test fields
During FUAVE project, the Sjökulla area will be further developed for the drone use. We are also constantly improving our portable reference targets.
If you need to test your own system, the Sjökulla test field in Uusimaa region in southern Finland is open for you! If you are interested, please contact us. During 2021–2022 we will also organize FUAVE events, where drone professionals are invited to test their systems in the Sjökulla or in temporal test fields.
To get more information, join the FUAVE newsletter and indicate your interest to Sjökulla test field in the FUAVE website’s Contact section.
Contacts
Eija Honkavaara eija.honkavaara@nls.fi
Eero Ahokas eero.ahokas@nls.fi
Lauri Markelin lauri.markelin@nls.fi
Juha Suomalainen juha.suomalainen@nls.fi