Part 2 – Generating 3D Maps

Chapter 2.2 – Digital Elevation Models

Overview

Once the 3D dense point cloud is generated, the next step is to create an accurate digital elevation model (DEM), digital surface model (DSM) or digital terrain model (DTM).

Before proceeding, be sure to save the project.

Building a DEM

A digital elevation model is a raster image (pixels, arranged in rows and columns) with each pixel representing an elevation. If the model represents elevations of the ground along with heights of structures such as buildings and trees, it can be referred to as a digital surface model (DSM). If the elevations only represent the ground or the bare earth, the model is referred to as a digital terrain model (DTM). The general term describing any type of raster elevation surface is digital elevation model (DEM).

From the Workflow menu, select Build DEM…. The dialogue box settings are described below.

Projection Type can be set according to other data sets that you’d like this DEM to geographically coincide with.

Source data allows a choice between using the Sparse CloudDepth Maps, or Dense Cloud as the source of elevation data. The Sparse Cloud should never be used as it is the least accurate representation of elevation. The main difference between Depth maps (if available) or Dense Cloud is that the latter is the truest representation of the surface but may have holes or gaps where elevation data may be missing due to sparse alignment between photos. The former represents a more continuous surface (assuming that interpolation or extrapolation was enabled when building the mesh), keeping in mind that the interpolated elevations are algorithmic guesses based on surrounding known elevations.

Interpolation can be Enabled (default) to fill holes or gaps in elevations, Disabled or an Extrapolation can be applied instead.

Point classes can be selected (if they have been previously created through classification) or All (default) classes will be used.

Region settings can be adjusted if necessary, affecting the output Resolution and Total size of the DEM.

Fig. 16 below shows the suggested DEM settings for this project and don’t worry if the Region values are slightly different.

Screen shot from MetaShape showing default settings in the Build DEM dialogue box including projection set to geographic and EPSG 6653, source data is dense cloud, interpolation is enabled,
Fig. 16 Build DEM parameters.

To view the DEM, double click it in the Workspace pane which brings up a new Ortho tab in the main view.

Screen shot from MetaShape showing the results of the generation of a DEM showing elevations between -3.29m to 30.2m
Fig. 17 Digital Surface Model (DSM) for Band 1 (Blue).

Note that the pixel resolution of the DSM in Fig. 17 is approximately 12.3cm. Also notice the clear demarcation of the lower and higher areas of vegetation.

The alternate type of elevations surface that can be created is a digital terrain model (DTM). A DTM shows a statistical approximation of the ground or bare earth surface under vegetation and with all surface features (i.e. buildings and trees) filtered out. To build a DTM, the first step is to classify the dense point cloud.

From the Tools menu, select Dense Cloud and then select Classify Ground Points….

Configure the parameters of the Classify Ground Points window as per Fig. 18 in order to generate three classes of points; Created (Never Classified), Ground and Low Point (Noise).

Screen shot of MetaShape showing the Classify Ground Points dialogue box with max angle set to three degrees, max distance set to 0.01m and cell size set to 40
Fig. 18 Classify Ground Points parameters.

The next step is repeat the creation of the DEM, but this time specifying which class (Ground) to generate the elevation model with (Fig. 19).

Screen shot of MetaShape showing the resulting digital terrain model with elevations ranging between -2.5 to 4.4m
Fig. 19 Results of the classification of the dense point cloud showing the digital terrain model (DTM).

Now that an accurate raster representation of the surface is generated, it can be used to create an orthophoto mosaic as described in the next Chapter.

License

Icon for the Creative Commons Attribution-ShareAlike 4.0 International License

Processing Multi-spectral Imagery with Agisoft MetaShape Pro Copyright © 2020 by Dr. Eric Saczuk is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License, except where otherwise noted.

Share This Book