What is a geo-referencing process?
After having performed a proper mapping of the environment you can perform the geo-referencing process. The geo-referencing process is a one-time process that allows Dragonfly to determine the link between the features inside the 3D map and the real world distances and it is done by telling Dragonfly what are the features seen when the camera is placed at some reference points using a DWG file of the area of interest.
At the end of the geo-referencing process, Dragonfly will be ready to be used in production and it will provide the positions as:
- XY coordinates (in meters).
- altitude from the ground (in meters).
How to perform the geo-referencing step by step
Step 1 – Choose the equipment for the geo-referencing
Because of the fact that the moving asset might be cumbersome, it might become difficult to ensure that the camera (mounted on the moving asset) is accurately positioned over the spots that will have to be chosen for geo-referencing process! This is why we suggest to do the geo-referencing process with the camera attached firmly to an extendable stick that is attached to a trolley with big and good wheels (on which you can carry around a laptop running the Dragonfly App). Doing so:
- the extendable stick will allow you to set the camera at the same altitude at which it will be used in production (which is the same height at which you have created your maps).
- the small size of the trolley will allow you to accurately position the trolley (and thus the camera) over the reference points chosen for the geo-referencing process!
Step 2 – Plan the number and the locations of the virtual markers
To plan the placement of the virtual markers you have to keep in mind these important information:
- The minimum number of markers is 3. After the placement of the 3rd marker, Dragonfly will start providing the XY accurate coordinates of the camera. Nonetheless, you can place also additional markers if the area to be mapped is more complex than a simple square or rectangle.
- Regardless of the total number of markers set for your venue, in order to compute the position, Dragonfly uses the 3 markers closest to the estimated position of the camera in real-time.
- Dragonfly provides the highest accuracy within the virtual polygon created by connecting the 3 markers closest to the estimated position of the camera. So, if your position falls outside any of the virtual polygons created by connecting 3 markers closest to the estimated position, then that position might be a little bit less accurate.
- The markers should be placed in such a way that they are not in a row to each other thereby creating flattened virtual triangles.
Examples of a good and bad choice for placement of 3 markers
Step 3 – Set the markers
The markers can be set by entering the metric coordinates of each marker taken from a DWG of the venue inside the Dragonfly Web User Interface.
The process to define markers is the following:
- Start Dragonfly.
- Activate the Navigation only checkbox after having clicked on the 3 bars button.
- If the map of interest is not loaded then click on the Maps menu button and load the map of interest. If you don’t have a map then it means that you have skipped this page!
- Press on the green START button to start the navigation. The video feed should be displayed inside the video preview window. You can drag and drop the corner to modify the size of the video preview window.
- If nothing is shown inside the video preview then it means that the calibration file has not been selected OR the USB camera is not properly connected.
- The status of the Dragonfly engine will change from MAP LOADING and then LOST.
- Move your camera to start detecting the features associated to the shapes and objects in your environment.
- By now the status will switch from LOST to NAVIGATION. During the process described in the next lines, if the status switches from NAVIGATION to LOST you have to move back to the area already mapped until the status switches back to NAVIGATION and then you can keep following the process outlined.
- Move with the camera towards the first remarkable location where you will collect the first virtual marker. A remarkable location is a location inside your venue that can be identified without doubts in the reality (e.g. some marks on the grounds) and inside the DWG file of the venue. When the location is reached:
- open the DWG file with the metric coordinates of your venue (you can use any CAD application like DWG TrueView).
- retrieve from the DWG file the metric coordinates of the current location of the camera (using the MEASURE > LOCATE POINT button in the top menu of DWG TrueView).
- click on the Markers tab on the bottom bar.
- make sure the status is still NAVIGATION and then click on the Add current location button.
- The Add current location button won’t be visible if you are NOT in NAVIGATION!
- enter the X,Y metric coordinates taken from the DWG and the current real altitude of the camera (measured with a laser, for example).
- The marker is automatically saved within the map file and visible inside the JSON coupled with the BIN file of the map.
- Move with the camera towards the second remarkable location where you will collect the second marker. When the location is reached:
- open the DWG file with the metric coordinates of your venue.
- retrieve from the DWG file the metric coordinates of the current location of the camera (using the MEASURE > LOCATE POINT button in the top menu of DWG TrueView).
- click on the Markers tab on the bottom bar.
- make sure the status is still NAVIGATION and then click on the Add current location button.
- The Add current location button won’t be visible if you are NOT in NAVIGATION!
- enter the X,Y metric coordinates taken from the DWG and the current real altitude of the camera (measured with a laser, for example).
- The marker is automatically saved within the map file and visible inside the JSON coupled with the BIN file of the map.
- Move with the camera towards the third remarkable location where you will collect the third marker. When the location is reached:
- open the DWG file with the metric coordinates of your venue.
- retrieve from the DWG file the metric coordinates of the current location of the camera (using the MEASURE > LOCATE POINT button in the top menu of DWG TrueView).
- click on the Markers tab on the bottom bar.
- make sure the status is still NAVIGATION and then click on the Add current location button.
- The Add current location button won’t be visible if you are NOT in NAVIGATION!
- enter the X,Y metric coordinates taken from the DWG and the current real altitude of the camera (measured with a laser, for example).
- The marker is automatically saved within the map file and visible inside the JSON coupled with the BIN file of the map.
- You have now placed 3 virtual markers and so your location will be displayed as X-Y metric coordinates in the bottom-right area of the Dragonfly App!
- Move with the camera towards the first remarkable location where you will collect the first virtual marker. A remarkable location is a location inside your venue that can be identified without doubts in the reality (e.g. some marks on the grounds) and inside the DWG file of the venue. When the location is reached:
If the accuracy is not good in some specific spots then you can collect additional markers following the steps above keeping in mind the information inside Step 2!
If you want to remove or modify the X-Y coordinates any of the markers placed, you can do so by using the list button inside the Markers tab.
Please send us the information listed inside this page.
Information inside the Dragonfly UI
Central area
A 3D space where:
- the blinking arrow represents your camera. You can use your mouse to zoom.
- there is a floor plan (if available).
Top-left corner
- name of the current map.
Top-right corner
- the video preview on the right that shows what is seen by the camera where the dots represent the features associated to objects/shapes detected in your environment.
Bottom-right corner
- tracking FPS – provides an indication about the speed at which the system runs (it must be higher than 20 FPS to produce good results).
- map memory usage – weight of the map.
- Landmarks – number of features associated to objects/shapes detected in your environment and part of the current map.
- Keyframes – number of keyframes detected so far.
- Markers – number of ArUco markers detected so far.
- distance, driving time, average speed, stop time – usage information (provided after the placement of the third marker).
- X, Y, Z – these are the metric coordinates!
- roll, pitch, yaw – orientation in degrees of the device (more info inside this page).
Bottom-left corner
A toolbar is shown that can be used to manage the floor plan, view, devices, geofence and markers currently associated with the loaded map. Under the 3 lines menu in the most right side of the bottom bar you can click on:
- Show path – to hide/show the real-time path representing the movements of your camera.
- Show history path – to hide/show the historical path followed by your camera.
- Show device name – to hide/show the name of your device.
- Show markers – to hide/show the markers.
- Show triangle – to hide/show the lines connecting the markers.
- Show heatmap – to hide/show the heatmap of your device
Top bar
- Start/stop button – to start and stop the positioning engine to perform the mapping and then the geo-referencing process.
- Pause button – to pause the mapping process without stopping the engine.
- the 3D view button – to see the 3D map with the features detected (only visible when the engine is up and running!).
- the Position button – to see the 2D location of the device (over imposed on a floor plan image if available).
- the Search history button – see description inside this page.
- the Snapshot button – to save a snapshot of what is currently seen by the camera (only visible when the engine is up and running!).
- the down arrow – to visualize:
- the built-in parameters of the location engine.
- the camera parameters (that can be changed by clicking on the 3 lines button and then on the Camera Settings button).
- the 3 lines menu with these buttons:
- Navigation only – enable or disable the Navigation mode. While the system is in Navigation the map used for Navigation is not updated or changed during the navigation process. The navigation mode allows users to manage the virtual markers used for the geo-referencing of the map.
- Show image – which allows to hide and show the video preview available on the top-right side of the GUI.
- Enable Aruco – to enable/disable the usage of ArUco markers. The ArUco markers are artificial features used to make the maps more robust/reliable in areas where there is a low number or variability of the natural features. More info inside this page.
- Enable temporal mapping – this feature allows Dragonfly to use the features seen inside unexplored areas (areas which features are not within the production map file used) to keep providing a position without getting LOST.
- Store source feed video – which allows to save the original/raw video coming from the camera.
- Store processed video – which allows to save the processed video.
- Change next video file name – this feature allows Dragonfly to append a specific TEXT at the end of the next video file saved.
- Enable data store – to enable the storage of the location data.
- Save path in device – to enable the storage of the location data locally.
- Sync path with cloud – to enable the storage of the location data on the cloud.
- Pull all floorplans from the cloud – to download all floor plan metadata from the Dragonfly Cloud Server.
- Push all floorplans to cloud – to upload all floor plan metadata to the Dragonfly Cloud Server.
- Calibrate camera – to calibrate the camera as described in this page.
- Camera settings – to access all the settings related to your camera (shown ONLY when the location engine is STOPPED).