Parameters.md

Parameters for Projection Procedure

The dictionary created for the Projection Procedure contains the following information:

• directories to input files (image, DEM, GCPs).
• projection parameters.
  
  

Input for Projection

georef_webcam searches automatically through all subdirectories of in_dir to find the needed files.

Image file

An RGB image is required. So far .jpg & .png files have been tested & projected successfully.

DEM file

PRACTISE requires a DEM in cartesian coordinates in the ASCII-Grid format. If this is not the case, georef_webcam generates this data from the provided input. If the DEM needs to be projected to a cartesian Coordinate Reference System (CRS), the EPSG number of it is asked from the user.
Additionally, the spatial resolution of the DEM can be resampled. We recommend to do so (e.g to 50 or 100m), if a large area is covered by the image and the projection parameters are not accurately known. It is faster to run the projection estimations on the DEM with lower spatial resolution. As soon as the projection accuracy is satisfying, you can change the DEM path variable in the dictionary to the one with full resolution. Use the option to edit the dictionary for that (see here).

GCP file (optional)

Ground Control Points (GCPs) can be used to optimize the projection with the Dynamically Dimensioned Search (DDS) optimization, which is incorporated into PRACTISE. See Example for the required txt-file formatting.

Projection Parameters

After inserting the directories to the required input, the parameters for the projection procedure are collected. Those parameters can be divided in 4 different groups:

• camera position
• camera orientation
• sensor parameters
• additional parameters

Table with overview over all parameters

Parameter Type	mandatory for PRACTISE	additional input variables		aux variables
camera position	camera_DEMCRS_E camera_DEMCRS_N camera_offset	camera_epsg camera_Easting/camera_longitude camera_Northing/camera_latitude camera_DEMCRS_Z		camera_x camera_y camera_DEM_height
camera orientation	target_DEMCRS_E target_DEMCRS_N target_offset roll_angle_deg	Target coordinate approach: target_epsg target_Easting/ target_longitude target_Northing/ target_latitude target_DEMCRS_Z height_diff_target_cam	Orientation angle approach: yaw_angle_deg pitch_angle_deg	target_x target_y target_DEM_height distance
sensor parameters	focalLen_mm sensor_width_mm sensor_height_mm
additional parameters	buffer_around_camera_m

Planar Camera Position

At first the coordinate position of the camera has to be entered. You can enter the coordinates in any CRS. They will be stored in the following variables:

• camera_Easting/camera_longitude
• camera_Northing/camera_latitude

Additionally, the EPSG-Nr. of the CRS (camera_epsg) is collected. Then, georef_webcam automatically projects the camera location to the CRS of the DEM:
camera_DEMCRS_E, camera_DEMCRS_N

Planar Camera Orientation

Next, the horizontal orientation is defined. PRACTISE requires a target point for its projection.
The target point is defined as the geographic position of the landscape feature in the center of the image (see also Figure 2).
However, georef_webcam provides two ways to define the horizontal orientation:

• either by inserting directly the coordinates of a target point in the same way as the camera position.
• or by inserting a yaw angle in degree: yaw_angle_deg.
  Based on this angle, the required target point is created artificially at a distance of 1km to the camera.

The definition of the planar camera position and orientation parameters are depicted here:

_{Figure 1: Planar Camera Position and Orientation Parameters.}

Vertical Camera Position

Two options exist to define the vertical location of the camera:

• relative height of camera above ground / above DEM: camera_offset
• absolute height of camera: camera_DEMCRS_Z

To support the user in selecting a reasonable height value, the procedure tells you the DEM height at the camera location (camera_DEM_height) before.

Vertical Camera Orientation

Consecutively, you can define the vertical orientation using one of the following ways:

• by inserting a pitch angle in degree: pitch_angle_deg
• by defining the height difference between target and camera: height_diff_target_cam
• by setting the relative height of the target point above ground / above DEM: target_offset
• or defining the absolute target point height: camera_DEMCRS_Z

For both, vertical position and orientation, all other parameters are derived from the given input. The definition of the vertical camera position and orientation parameters are depicted here:

_{Figure 2: Vertical Camera Position and Orientation.}

Adjusting further parameters

All other parameters are set to default values, which however might not be correct. If values are unknown, keep default for a first try. But change them, if you know them. The predefined parameters are:

• Roll angle of the camera: roll_angle_deg (default: 0°)
• Sensor parameters given in mm:
  focalLen_mm (14 mm)
  sensor_width_mm (22.3 mm)
  sensor_height_mm (14.9 mm)
• Buffer range: buffer_around_camera_m (100m)
  This buffer variable is implemented in PRACTISE to make close range topography transparent for viewshed calculation and to avoid, that close slopes wrongly obscure the panoramic view.

Sensor size parameters (width & height) can possibly be found on digicamdb.com.

Execution of Projection

After collecting all required variables, the parameters are passed to PRACTISE and the projection is executed (see Readme).