Trimble GNSS and its combination with DJI Enterprise

 

The contribution of GNSS in the geospatial field is well known, probably consolidating itself as the most important geospatial observation technique that exists today due to its transversality. On the other hand, drones revolutionized a variety of fields, of which, without a doubt, one of the most outstanding is the geospatial area. Together, GNSS and drones, whether with photogrammetric or LiDAR techniques, stand as fundamental techniques to address any project involving topographic representations. Now, the important thing is to conveniently combine techniques, leveraging the benefits of productivity, precision, and representativeness resulting from their application.

In this sense, Trimble GNSS and software are prepared to be combined directly with DJI drones or to process their data, obtaining results with direct application in the industry.

 

A brief summary of the combination of both techniques

Mainly, GNSS is a discrete geospatial observation technique. GNSS determines the coordinates of points for a variety of applications. Often, GNSS involves densifying the geodetic reference framework on which topographic representations made from a drone will be produced. This is how GNSS has been used to determine the coordinates of control points that a block of photographs needs after aerial triangulation to georeference the representation. This technique is known as photogrammetry with indirect referencing, and GNSS has been fundamental in this process. In this case, GNSS can be used in RTK mode, using a base with known coordinates and obtaining the coordinates of the control points in the appropriate system (even in local systems). A modification of this technique involves the use of Trimble DA2, from Trimble RTX in real time, which increases productivity by not requiring a base station, achieving precisions similar to RTK.

Furthermore, GNSS contributes to determining the trajectory of a drone or any mobile object. For example, a drone equipped with a survey-grade GNSS can store satellite observations to precisely process its trajectory from a base station. This technique has been used in intensive drone aerial photogrammetry operations, achieving great productivity given that no ground control points are required. By obtaining the drone's trajectory, each position is individualized over time, allowing the interpolation of the position of a captured photograph and applying it directly in aerial triangulation, with the condition that the position has much greater precision than the orientation. This technique is known as direct georeferencing and is the first step to understanding the incorporation of inertial sensors. It should be noted that the direct georeferencing technique can be performed both in post-processing and in RTK, with interpolation being necessary in the case of post-processed trajectories, while in RTK, each photograph is directly geotagged.

If the drone incorporates a photographic camera, the position and orientation of each photograph must be considered to produce the final photogrammetry deliverables: point clouds and orthophotos. This is precisely what is achieved with aerial triangulation and dense reconstruction. If the camera has an event marker, conveniently articulated with a GNSS + IMU, it is possible to determine both the position and orientation of each photograph. However, this implementation can be somewhat expensive for photogrammetry applications in some projects. For this reason, the direct georeferencing technique has been so successful over the last decade.

In the case of having another sensor on the drone, such as a LiDAR, it is necessary to process the complete trajectory with GNSS + IMU, obtaining position and orientation for each instant. This means that after processing the trajectory, a complete record of each position and orientation of the sensor on board the drone is available. With this information, the sensor reproduces its pose with a defined periodicity, allowing the calculation of the resulting coordinates from the observations made.

 

Experience using RTK

A photogrammetric data acquisition experience is carried out using a Mavic 3E, which is receiving differential corrections directly from a Trimble base. This connection can be established in two ways:

• Receiving differential corrections via internet: an NTRIP broadcasting corrections in the format supported by the drone is required. The drone's controller always remains connected to the internet, and through telemetry, the system manages to geotag the photographs. This method is suitable when there is GNSS infrastructure, such as GEOCOM GNSS Network, and mobile internet connection.

• Receiving differential corrections via WiFi: a GNSS receiver like Trimble R12i is configured to transmit differential corrections via WiFi from an NTRIP, then the drone's controller receives them this way. This methodology is suggested when the drone needs to navigate based on a defined coordinate system or when there is no GNSS infrastructure and/or internet.

 

Figure 1: Mavic 3E - RTK connection

 

Once the link for the drone to receive differential corrections is configured, the desired flight plan is developed, considering that each photograph will have its precise position from the RTK link. In this case, a Mavic 3E was used, which has RTK navigation.

 

Figure 2: Mavic 3E - flight planning

 

For its part, GNSS establishes a direct link to the drone for photo tagging and thus performs aerial triangulation. Additionally, a series of profiles are determined with GNSS RTK to directly compare with the point cloud, along with horizontal validation of some marks.

 

Figure 3: Trimble R12i

 

Comparison Results

The photographs and all data produced by Mavic 3E were imported directly into Trimble Business Center, which allows photogrammetric processing of a variety of drones, including those from DJI. Trimble Business Center provides a complete environment not only for producing photogrammetric deliverables but also for processing point clouds, drawing, and performing topographic calculations to enhance topographic representation.

 

 

Figure 4: Importing photogrammetric data into TBC

 

Regarding photogrammetric processing, since the photographs have precise positioning, it is not necessary to process the trajectory, which results in efficient and very simple processing. Only aerotriangulation is performed, which is adjusted to the reference framework given by the GNSS base, and finally, the production of final products.

 

Figure 5: Excerpt from the aerial photogrammetric processing report in TBC

 

In terms of the comparison made, two very simple experiments were conducted. The first involves comparing GNSS determination on materialized points that act as a photo-identifiable element with the corresponding position obtained from the orthophoto. 4 points are identified, yielding an average difference of 4.4 ± 1.9 cm.

 

Figure 6: Horizontal comparison

 

The second experiment relates to comparing the GNSS position with the corresponding position in the point cloud. Of course, this comparison allows for establishing a vertical criterion. After comparing 65 points, a mean value of 5.8 ± 4.1 cm vertical difference is obtained.

 

Figure 7: Vertical comparison

 

Conclusion

The georeferencing of photogrammetric products is a key aspect for their use in different applications, allowing for a unique reference that enables robust and precise multi-temporal analysis. In this sense, the georeferencing process must possess a high level of automation to reduce the time associated with the process itself, while ensuring accuracy in its results. As presented in this experience, thanks to the constant evolution of drones in the last decade, RTK direct referencing provides a reduction in the processing times of the drone's trajectory, which is performed simultaneously while flying without the need for additional processing. This is achievable either through a local network or through an NTRIP-based approach, offering an efficient, stable process and providing precisions compatible with those expected for the photogrammetric product(s).

Regarding the results obtained through photogrammetric processing, Trimble Business Center's Aerial Photogrammetry module stands out as intuitive software that generates representative and reliable deliverables that will serve as the basis for the subsequent generation of new products that enhance topographic representation. On this last point, it is key that TBC offers a complete environment with different modules (point cloud, surfaces, corridors, CAD, among others) that will adjust to the final product expected.

Regarding the validation of the obtained products for GNSS checking, it is important to ensure that the definition of the checking points is reliable. In this sense, the RTK mode for GNSS can be addressed through "topo point" occupations that provide greater confidence or also through the establishment of RTK networks.

 

Acknowledgements

GEOCOM thanks Rumbos Geomensura for their support in this experience.