ddm

Photogrammetry Tutorial

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).

Photogrammetry_PS_00.jpg

Getting Started

Other Advanced Tutorials

Part 1: Image Aquisition

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.

Photogrammetry_Pix4D_01.jpg

Photogrammetry_Pix4D_02.jpg

Please take note of the following points:

Part 2: Image Processing

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.

Creating a new project

Once you have your images on your local drive, open Agisoft Photoscan Professional and create a new project.

Photogrammetry_PS_01.jpg

Add the images

Photogrammetry_PS_02.jpg

Align the images

Photogrammetry_PS_04.jpg

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:

Photogrammetry_PS_05.png

Build a dense cloud

Photogrammetry_PS_06.jpg

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:

Photogrammetry_PS_07.png

Classify ground points

Photogrammetry_PS_13.jpg

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”:

Photogrammetry_PS_19.png

Build a Digital Terrain Model

Photogrammetry_PS_15.jpg

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.:

Photogrammetry_PS_20.png

Only Ground Points:

Photogrammetry_PS_22.png

Build an Orthophoto

Photogrammetry_PS_16.jpg

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:

Photogrammetry_PS_21.png

Part 3: Exporting Data

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.

Export a full point cloud

Photogrammetry_PS_12.jpg

Export ground points only

Photogrammetry_PS_14.jpg

Export the Digital Terrain Model

Photogrammetry_PS_17.jpg

Export the Orthophoto

Photogrammetry_PS_18.jpg

end

This site is open source. Improve this page.