The official site of
Digital Design Methods
Photogrammetry is the science and technology of obtaining reliable information about physical objects and the environment through the process of recording, measuring and interpreting photographic images and patterns of electromagnetic radiant imagery and other phenomena [wikipedia]. This tutorial will cover the basic steps of obtaining a Digital Terrian Model (DTM) from orthographic images aquired by a consumer quadcopter (without ground control points).
We will aquire the images using a consumer quadcopter and the Pix4D Capture App. Make sure you are aware of all up-to-date regulations regarding your flight-location and flight-mission using the links above. Before you fly a drone yourself, please review the following check-list that applies only for Switzerland, until 12.12.2020:
There are many apps that can do waypoint-missions for DJI (or other) drones like Pix4D Capture, DroneDeploy, QGroundControl, Litchi, etc. Be aware that not all drones are supported, and check compatibility before borrowing or buying a drone for aerial surveys. We will use Pix4D to capture the images.
- Once you open the app Pix4DCapture on your Android or Apple device, choose Double Grid Mission for best results.
- You can now select the area you want to survey, and press start to lauch the mission.
Please take note of the following points:
The processing of the images will be done using Agisoft Photoscan Professional (now called Agisoft Metashape). This is a stand-alone software product that performs photogrammetric processing of digital images and generates 3D spatial data to be used in your landscape architectural projects. Alternative software is available like Pix4D Mapper, Autodesk ReCap Photo, DroneDeploy, etc.
Once you have your images on your local drive, open Agisoft Photoscan Professional and create a new project.
- Save your project on a local drive
- Go to the menu item > Workflow and click > Add Photos
- Go to the menu item > Workflow and click > Align Photos
- Use a low accuracy to start, this process can take a long time…
- Pair Selection: ‘disabled’
At this stage PhotoScan finds matching points between overlapping images, estimates camera position for each photo and builds sparse point cloud model. You should now have a preview of the 3D model in low resolution, with the position and direction of the photos:
- Go to the menu item > Workflow and click > Build Dense Cloud
- Use a low quality to start, this process can take a long time…
Based on the estimated camera positions the program calculates depth information for each camera to be combined into a single dense point cloud. You should now have the full resulution of the point cloud:
- Go to the menu item > Tools > Dense Cloud > Classify Ground Points
- From: Any Class
- In order to get the best results, you might have to change the angle, distance, and cell size parameters. The following parameters worked well for me:
- Max angle (deg): 15
- Max distance (m): 0.5
- Cell size (m): 25
To create a Digital Terrain Model in the next step, this step differentiates ground-points from other elements (like trees, buildings, etc). You can view the classified points by clicking on the menu bar icon “Dense Cloud Classes”:
- Go to the menu item > Workflow and click > Build DEM
- Coordinate System: CH1903/LV03
- Source Data: Dense Cloud
- Interpolation: Enabled
- Point classes: Ground
You can view the Digital Elevation Model by double-clicking on the DEM-layer in the Workspace Pane:
Digital Surface Model including trees, buildings etc.:
Only Ground Points:
- Go to the menu item > Workflow and click > Build Orthomosaic
- Coordinate System: CH1903/LV03
- Source Data: Dense Cloud
- Interpolation: Enabled
- Point classes: Ground
Note: If you are not happy with the result, try building a mesh surface first, and then re-build the orthophoto from the mesh instead of the DEM. You can view the Orthophoto by double-clicking on the Orthomasaic-layer in the Workspace Pane:
Please note: Because we have not registered the images with ground control points, the data will not be in its exact geographical location, e.g. likely a few meters off in x and y axis. It will also not have the correct z-heigth. After you export your data you can view it in CloudCompare/GIS.
- Go to the menu item > File and click > Export Points
- Select ASTM E57 (.e57)* as export format.
- Coordinate System: CH1903/LV03
- Source Data: Dense Cloud
- Point Classes: All
- Point Colors: [x]
- Split in blocks: []
- Go to the menu item > File and click > Export Points
- Select ASTM E57 (.e57)* as export format.
- Coordinate System: CH1903/LV03
- Source Data: Dense Cloud
- Point Classes: Ground
- Point Colors: [x]
- Split in blocks: []
- Go to the menu item > File > Export DEM > and click Export TIFF/BIL/XYZ
- Projection: CH1903/LV03
- Click Export
- Save as XYZ (*.xyz)
- Go to the menu item > File > Export Orthomosiac > and click Export JPEG/TIFF/PNG
- Projection: CH1903/LV03
- Click Export
- Save as JPEG/PNG, whatever you prefer(the TIFF format maintains the geographic location when you open it in a GIS program).
end