8.3 Software for Drones

There are several types of software that may be required for different types of drones and use cases. Here are a few examples:

Flight planning and control software This type of software is typically used to plan and control drone flights, it allows you to set waypoints, plan routes, and monitor the drone’s flight status in real-time. Some examples include DJI GO, DJI Ground Station Pro, and Pix4Dcapture.

Photogrammetry Software - This type of software is used to process the images captured by the drone and create maps, 3D models, and point clouds. Examples include Agisoft Metashape, DroneDeploy, and WebODM.

Ground control systems - Ground control systems, such as GNSS Base Stations are typically used for survey-grade drones, as it helps to improve the accuracy of the images captured by the drone, as well as process and analyze the data.

Geospatial Analysis Software - Once you have processed data from a drone, the next step is to analyze the data, commonly in within geographic information system (GIS) software. GIS software can be used for a variety of analyses such as 3D modeling, change detection, image analysis and surface analysis.

8.3.1 Photogrammetry Software

The choice between DroneDeploy, Agisoft Metashape, and Pix4D will depend on your specific needs and goals.

DroneDeploy is a cloud-based software that is easy to use and well-suited for creating detailed 2D and 3D maps and models of large areas, such as construction sites, mines, and farmland. It also includes features such as automatic flight planning and easy sharing of maps and models.

Agisoft Metashape is a photogrammetry software that is used for creating detailed 3D models, orthomosaics, and point clouds from drone-captured images. It is more advanced than DroneDeploy and it is intended for professional users such as surveyors, engineers, and architects. It offers more control over the processing and it can handle larger datasets and more complex projects.

Pix4D offers a complete solution for processing drone data and creating maps, 3D models, and point clouds. It can be used for a wide range of applications such as surveying, agriculture, and construction. Pix4D is known for its high accuracy, advanced features, and the ability to handle large datasets.

If you are a professional user that needs high accuracy and advanced features and have experience with photogrammetry software, Agisoft Metashape or Pix4D might be the best choice. If you are a beginner, or you have a specific use case such as creating maps for agriculture or construction, DroneDeploy might be the best choice for you.

WebODM is an open-source software for processing drone-captured images and creating 2D and 3D maps, models and point clouds. It is built on top of the OpenDroneMap (ODM) library and provides a web interface for running ODM on a remote server. However, it is not as advanced as the other commercial softwares.

8.3.2 Neural Reference Fields and Gaussian Splats

Recent advances in 3D scene reconstruction from drone imagery have been driven by machine learning techniques such as Neural Radiance Fields (NeRFs) and 3D Gaussian Splatting. NeRFs use neural networks to represent a scene as a continuous volumetric function, allowing photo-realistic novel views to be rendered from a sparse set of input images. This technique excels at capturing fine-grained lighting and geometric detail, but can require extensive training time and compute resources.

Gaussian Splatting is a newer, more efficient alternative that represents the scene using a collection of 3D Gaussians with color, opacity, and size attributes. Rather than relying on neural networks to synthesize views, Gaussian Splatting renders directly using rasterization, enabling real-time performance and faster reconstruction. For drone users mapping landscapes or structures, Gaussian Splatting offers a practical path to high-fidelity 3D visualization with minimal post-processing and can be generated from videos, rather than mapping flight patterns making it a potentially valuable tool where visualization of a 3D environment is more important than 3D modeling, such as in educational outreach activities.

Both techniques are reshaping workflows for inspection, mapping, and environmental modeling - while experimental, they may become valuable tools in the modern drone software stack.

Nerfstudio – Nerfstudio Modular NeRF pipeline with drone and GPS metadata support.

Instant-NGP – Instant-NGP Fast local NeRF rendering from NVIDIA.

Luma AI – Luma AI Mobile-friendly app for quick NeRF capture and sharing.

Poly.cam - Poly.cam AI powered reality capture for cloud-based photogrammetry and gaussian splat generation.