Agisoft Photogrammetric Software: Basic Training - Post Processing

Align Photos

Once photos are loaded into Photogrammetric Kit, they need to be aligned.  At this stage, the software finds the camera position and orientation for each photo, and builds a sparse point cloud model.

  • To align a set of photos:
    • Select the Align Photos… option from the Workflow menu.
    • In the Align Photos dialog box, select the desired alignment options and then click OK.

After some time, the alignment will finish and display a sparse point cloud model in the Model view.  Camera positions and orientation are indicated by blue rectangles.

The photo positions from the Falcon 8 are derived from uncorrected GPS positions.  In some cases this can be an accuracy of 10m.  In order to improve the absolute accuracy, it is important to tie the data to ground control points

Build Mesh

Markers are used to optimize camera positions and orientation data, which provides better model reconstruction results and precise geo-referencing.

  • To be able to follow a guided marker placement approach (which is faster and easier) it is recommended to first reconstruct geometry.
    • Select Build Mesh… from the Workflow menu and fill out the parameters in the Build Mesh dialog.

Set Projection

Building the geometry typically takes under a minute to reconstruct for the sparse point cloud.  Once this has completed, the GCPs can be imported. 

  • Start by setting the coordinate system in which the ground control points are projected.
  • Make sure all images are checked. Click the Select Projection button, seen in diagram outlined in red. Next select the corresponding coordinate system for the GCPs.

Import Markers

  • Import the marker coordinates from a file by clicking the Import button located on the Reference pane (seen in diagram outlined in red) and select the file which contains the GCP coordinate data.
  • After selecting to import markers, a dialog box will appear where the user will need to indicate the delimiter according to the structure of the file. Select the row to start loading from.  Once set, click “OK”.

Match Markers

Once the Markers are imported, the next step is to match them with the GCPs which can be seen from the images taken from the Falcon 8.

  • Under the Reference tab, Right click on one of the markers and select the option Filter Photos by Marker.
  • The photos will be filtered by location of a selected marker so that all photos within the range of this marker are shown.
  • Double clicking on one of the filtered photos allows the user to check the marker location and refine its position as necessary to provide maximum accuracy. 
  • Open each photo where the created marker is visible, zoom in and drag the marker to the correct location while holding left mouse button.  Repeat the described step for every GCP.
  • To achieve higher accuracy in calculating the camera parameters and to correct possible distortion, an optimization procedure should be run.  Click on the Settings button (shown in diagram outlined in red) in the Reference pane.
  • Set the desired values for the parameters in the measurement accuracy section and click OK once finished.
    • Camera accuracy: 10
    • Marker accuracy: 0.005 ​(It is recommended that the marker be set at 0.005 value if the real marker accuracy is within 0.02m).
  • On the Reference pane, uncheck all photos and check the marker(s) to be used in the optimization procedure. The markers which are unchecked are not taken into account for data transformation and can serve as check points to evaluate the optimization results. 
  • Click Optimize button (show in diagram outlined in red) on the Reference pane toolbar.  Select camera parameters to be used in optimization and click OK to start the process and apply the Ground Control Points.


Build Cloud

Once the matching and optimization has been completed, the next step is to build the dense point cloud.

  • Select the Build Dense Cloud command from the Workflow menu. 
  • Set the following recommended values for the parameters in the Build Dense Cloud dialog:
    • Quality – Medium/High
    • Depth Filtering – Aggressive

Build Mesh

After dense point cloud has been reconstructed, it is possible to generate polygonal mesh model based on the dense cloud data.

  • Select the Build Mesh command from the Workflow menu. 
  • Set the following recommended values for the parameters in the Build Mesh dialog:
    • Surface type – Height Field
    • Source data – Dense Cloud
    • Polygon Count – Medium
    • Interpolation - Enabled

Build Orthomosaic

After Mesh has been reconstructed, it is possible to generate Orthomosaic model based on the dense cloud data.

  • Select the Orthomosaic command from the Workflow menu. 
  • Set the following:
    • Projection type – Geographic
    • Projection – by default the projection set in the Reference Settings is used
    • Surface – Mesh
    • Blending mode - Mosaic