 |
Overview
Photogrammetry is a technique used to derive accurate measurements and create detailed 3D models of real-world objects or landscapes by utilizing multiple photographs. The advent of Unmanned Aerial System (UAS) technology has significantly increased the importance of photogrammetry, especially when executed correctly with specialized photogrammetry software. The aim of this post is to review essential photogrammetry terms that are commonly used in Pix4D, but can also be applied to general UAS photogrammetric operations.
Photo Overlap
Referecning Figure 1, photo overlap is the amount of shared area between adjacent images captured during a drone flight plan. Adequate photo overlap is crucial for ensuring that photogrammetry software can accurately stitch images together. Generally, photo overlaps between 60% and 90% are considered suitable. However, the ideal value can vary depending on the type of area being captured. For instance, when using the Zenmuse X/7 sensor from DJI, I typically aim for an 80% overlap especially since most of the envorinments that I capture include moderate to heavy vegetation. It's important to note that high photo overlap may become less effective and more time to process, but this can also depend on the specific circumstances.
 |
| Figure 2:Example of Photo Overlap |
Geometry
Geometry is the study of shapes, sizes, and positions of objects in space. In photogrammetry, geometry plays a crucial role in determining items such as camera gocal length, image orientation, image mtching, triangulation, surface reconstruction, and coordinate system transformation . Drone images captured at different perspectives are used to reconstruct the geometry of the scene, ultimately producing a 3D model. Referencing Figure 3 is an example of a colinear characteristics from geometry.
 |
| Figure 3: Colinear characteristics from geometry |
Radiometry
Radiometry is the science of measuring electromagnetic radiation, including visible light. In photogrammetry, radiometry is essential for capturing accurate color and brightness information from images. The radiometric quality of an image affects the accuracy and quality of the final photogrammetric product. Radiometric corrections are used to remove inconsistencies caused by variations in the intensity of light, atmospheric conditions, and other factors that can affect the quality of the images. This is particularly important when using photogrammetry to create 3D models or maps of large areas because any inconsistency in the images can result in errors in the final product. See Figure 4 for a radiometry diagram
 |
| Figure 4: Radiometry Diagram |
Triangulation
Referenced in Figure 5, triangulation is a process in which the position of a point is determined by measuring angles from known reference points. In photogrammetry, triangulation is used to determine the position of objects in 3D space using overlapping images. This process is crucial for creating accurate and detailed 3D models and maps.
 |
| Figure 5: Triangulation |
Internal and External Parameters
Depicted in Figure 6, internal parameters refer to the characteristics of the camera used to capture images, such as focal length, sensor size, and lens distortion. External parameters include the position and orientation of the camera relative to the object being photographed. Both internal and external parameters are crucial for accurate photogrammetric processing.
 |
| Figure 6:Internal Pararmaters |
Initial and Computed Parameters
Initial parameters are the approximate values of internal and external parameters, typically estimated using metadata from the drone and camera. Computed parameters are the refined values obtained through photogrammetric processing. Accurate computed parameters are essential for generating high-quality 3D models and maps.
Real-Time Kinematic
(RTK) is a GPS-based technology for UAS (Unmanned Aerial System) operations. By utilizing a network of reference stations, it corrects the GPS signal errors in real time, enhancing positional accuracy to centimeter levels. RTK enables precise navigation, improves flight stability, and increases the efficiency of UAS operation. The technology also allows for safer and more reliable operations in complex environments. RTK typically costs more compared to traditional GPS systems. Click on the video below for a video summarizing RTK.
Post-Processed Kinematic
PPK, combines GPS data from both the drone and a base station, enabling precise determination of the drone's position during flight. The collected data is processed after the flight, correcting any errors and refining positional information. This results in high-precision maps and 3D models,. PPK's main advantage is its ability to deliver centimeter-level accuracy without requiring real-time data transmission.
Coordinate System
A coordinate system is a standardized method of representing the position of points in space. In UAS related mapping, coordinate systems are essential for defining the location and orientation of objects and images accurately. The Universal Transverse Mercator (UTM) and the World Geodetic System (WGS84) are commonly used coordinate systems in photogrammetry. The specific coordinate system used may vary depending on the location, but typically, I use the state plane coordinate system unless there is a request for something else. To learn more about coordinate systems, click on the video below.
Tie Points
Tie points are identifiable features that are visible in multiple overlapping images and depicted in Figure 7 which was taken from a point cloud I created using Pix4D. Photogrammetry software uses tie points to establish relationships between images, align them accurately, and reconstruct the 3D geometry of the scene. The quality and quantity of tie points are essential for generating accurate and detailed 3D models and maps.
 |
| Figure 7: Looking at Tie Points Note Number of Images Accociated With Green Ray |
Ground Sampling Distance (GSD)
GSD is a measure of the distance between two consecutive pixel centers on the ground, representing the smallest object that can be distinguished in an image. Referenced in the video below at 9:30 lower GSD values indicate higher image resolution and detail. GSD is influenced by factors such as camera sensor size, focal length, and flight altitude.
Volume Measurement
A volume measurment used to estimate the volume of objects or features, such as stockpiles, excavations, or earthworks. By creating a 3D model of the object or area, photogrammetry software can accurately calculate the volume, providing valuable data for project planning and management. Figure 8 depictes a volume measurement taken by data I captured with the DJI M210 Zenmuse X/7.
.png) |
| Figure 8: Example of Volume Measurment of Point Cloud |
Absolute Accuracy
Absolute accuracy refers to the degree of closeness between the measurements or positions derived from photogrammetric processes and their true or actual values in the real world. In other words, it is a measure of how accurately the photogrammetric data, such as point coordinates or feature measurements, represents the true physical locations of the objects or features being measured.
Relative Accuracy
Relative accuracy measures the consistency of distances, angles, and positions between objects within the photogrammetric model. High relative accuracy is crucial for creating accurate 3D models and ensuring that the spatial relationships between objects are maintained. To reinforce the concept of absolute and relative accuracy, refere to the video below.
Structure from Motion
(SfM) is a photogrammetric technique that reconstructs 3D structures from a sequence of 2D images by analyzing the motion of the object or camera. It involves feature extraction, matching, camera pose estimation, and point cloud generation to create an accurate 3D mode. For a great explanation of structure from motion, visit the video below.
Root Mean Square Error
(RMSE) which is depicted in Figure 9, is a statistical measure used to quantify the average deviation between predicted and observed values. In the context of Unmanned Aerial Systems (UAS), it assesses the accuracy of measurements like position or elevation compared to ground truth data. Lower RMSE values indicate better accuracy, while higher values suggest larger discrepancies between the UAS measurements and the reference values.
 |
| Figiure 9: Note RMSE |
Ground Control Points
Referenced in Figure 10, (GCPs) are easily identifiable, well-defined, and accurately surveyed reference points on the Earth's surface that are used to improve the geolocation accuracy and precision of aerial imagery captured by drones or other remote sensing platforms.
 |
| Figure 10: Example GCP Targets |
Digital Elevation Model
(DEM) in UAS is a digital representation of the Earth's surface or terrain that is created using elevation data collected by aerial sensors mounted on unmanned aerial vehicles (UAVs), It displays the elevation information as a grid of cells where each cell corresponds to a specific area on the ground and has an associated elevation value. A DEM can include above ground features. See Figure 11 for details.
Digital Terrain Model
A DTM is a specific type of DEM that focuses on the bare ground surface, excluding above-ground features.
Digital Surface Model
A (DSM) in the context of refers to a 3D representation of the Earth's surface that captures both natural and man-made features, including buildings, vegetation, and terrain. See Figure 11 for details.
 |
| Figure 11:Difference Between DSM, and DTM |
Conclusion
By comprehending this list of terms, you will gain a deeper understanding of the quality of photogrammetric data derived from Unmanned Aerial System (UAS) technology. The terminology described is not limited to Pix4D users but is also relevant to general UAS photogrammetric operations. If you think there are any additional terms that should be included, please feel free to share your suggestions.
Citations
Pix4D. (2023, April 28). Ten Basic Terms of Photogrammetry Knowledge. Pix4D Blog. https://www.pix4d.com/blog/ten-basic-terms-photogrammetry-knowledge/
Plex.Earth Support. (n.d.). Elevation Modeling: the differences between DTM, DSM, & DEM. Retrieved April 28, 2023, from https://support.plexearth.com/hc/en-us/articles/4642425453201-Elevation-Modeling-the-differences-between-DTM-DSM-DEM
