Lab Overview
This lab introduced the use of Pix4D software to process UAS imagery into usable 3D and mapping products. The main focus was learning how to properly set up a project, check important settings, and understand the outputs generated during processing. The lab emphasized that while Pix4D is user friendly, incorrect settings can lead to poor results, making it important to understand what the software is doing at each stage.
The data processed in this lab came from previous Skydio 2+ flights and included two separate missions. The overall goal was to generate point clouds, 3D meshes, and become familiar with Pix4D’s help documentation and workflow.
Accessing Pix4D and Help Resources
The lab began by logging into Pix4D Mapper using the shared site license. Students were required to properly log out after use to avoid tying up licenses. Once logged in, the class explored Pix4D’s built in help resources, including the Video Academy and the Getting Started documentation.
These resources were used to answer questions related to coordinate systems, processing templates, camera models, and common data products such as contour lines and point clouds. This part of the lab reinforced the importance of using official documentation to troubleshoot and learn new workflows.
Project Setup and Data Organization
Before processing began, a structured folder system was created on the local machine to store collection, processing, and analysis data. Separate folders were made for each mission to keep datasets organized and easy to identify later.
Images from the Week 4 Skydio flights were placed into their respective folders. Before importing the data into Pix4D, image properties were reviewed to confirm camera information, exposure settings, and geotagging. The imagery was confirmed to be geotagged, meaning each photo contained latitude and longitude data.
Image Properties and Camera Settings
Once the images were added to Pix4D, the image properties window was reviewed. Sensor information was checked to ensure accuracy, especially due to the Skydio 2+ using a rolling shutter camera. The camera model was manually adjusted to use a linear rolling shutter setting to reduce distortion in the final products.
The image coordinate system and output coordinate system were reviewed but left at default settings for this lab. The correct processing template was selected to match the mission type.
Processing Workflow
Both missions were processed using the 3D Model template. All processing steps were enabled, including initial processing, point cloud generation, and mesh creation. Processing options were briefly explored to understand where additional outputs such as contour lines could be generated.
After processing completed, the quality and completeness of the outputs were reviewed. The densified point cloud and triangle mesh were generated for each mission, allowing for 3D visualization and inspection of the scanned objects.
Results and Outputs
The final outputs included densified point clouds and 3D triangle meshes for both the car and tower scans. These products allowed for detailed visualization of the structures and demonstrated how image quality, overlap, and camera settings affect reconstruction results.
Students navigated through the models to inspect detail and completeness. Animated fly through videos were also created using Pix4D’s tools to produce short 360 degree views of each model.
Figure 1: Car 3D Scan Point Cloud
Figure 2: Car 3D Scan Triangle Mesh
Loading video: 3d car scan animation.mp4...
Figure 3: Tower 3D Scan Point Cloud
Figure 4: Tower 3D Scan Triangle Mesh
Loading video: lightpole_week4_S2_flyaround.mp4...
Key Takeaway
This lab demonstrated the full workflow of processing UAS imagery in Pix4D, from organizing data to generating final 3D products. It highlighted the importance of checking camera models, understanding processing steps, and using help documentation when questions arise. Overall, the lab provided a strong foundation for future work involving ground control points, project merging, and more advanced Pix4D analysis.