Calibrating the virtual loop - SharpV | SharpOS 13.6

For parking applications where the license plate capture rate is critical, the SharpV virtual loop feature can detect vehicles with damaged or dirty license plates that are not detected by the SharpV ALPR camera. The plate numbers of these vehicles can then be manually modified in Security Desk.

1. Log on to the Sharp Portal.
2. From the Configuration menu, select the Analytics > Virtual loop page.
3. Click Enable.
   The camera saves the configuration and displays the message "Configuration saved successfully".
4. Click Configure.
   The status LED on the SharpV flashes red and green during the configuration.
5. Match the markers.
   1. In the virtual loop configuration page, the video feeds from the ALPR camera and context camera are displayed with an A and B marker for each camera. Choose two points that are visible in both camera feeds and move the A and B markers so that they are in the identical positions in each camera feed. When the markers are close to the correct position, an orange box appears in the context camera video feed. Fine tune the marker positioning so that the orange box matches the field of view of the ALPR camera.
      NOTE:

      • For best results, choose A and B marker locations that are as far apart as possible vertically and horizontally.
      • If either of the video feeds are too dark or too bright to accurately place the markers, you can temporarily adjust the brightness using the sliders located above each video feed. When the calibration is finished, the exposure returns to its default mode.
   2. Click Next.
      
6. Configure a detection zone.
   1. To configure the system to detect vehicles that are moving in a specific direction, select one of the following detection modes from the drop-down list:
      • Enters the field of view in the detection zone.
      • Exits the field of view in the detection zone.
      • Enters or exits the field of view in the detection zone.
   2. To draw a detection zone in the camera video feed, create a polygon by clicking on at least three points.
      NOTE:

      • Clicking in the field of view clears any existing polygon. You can also click Clear the zone to remove an existing polygon.
      • To draw a polygon that covers the entire context camera field of view, click Select all and use the Enters or exits the field of view in the detection zone setting.
      Consider the following when drawing the detection zone:

      • The polygon must touch the border of the image. This is necessary for all direction of travel settings, for example, Enters the field of view in the detection zone.
      • The polygon lines cannot intersect.
      • To be detected, at least 25% of the visible part of the vehicle must pass through the red detection zone (if configured) when the vehicle enters or exits the image. The vehicle must also cover at least 20% of the orange ALPR field of view somewhere along its trajectory.
      • Try to draw your detection zone so that cyclists and pedestrians do not pass through the zone. For example, if you have too many false detections when you draw a polygon far from the camera and use the Exits the field of view in the detection zone setting, try drawing a polygon close to the camera and use the Enters the field of view in the detection zone setting.
      For more information, see Virtual loop detection zone examples.
   3. When you have finished drawing the polygon, click Done.
      The system displays In progress - Waiting for 5 more vehicles and counts down the number of license plate reads that are required for calibration to finish. During this calibration, the system evaluates the expected vehicle size and trajectory.
      NOTE:

      • The system completes the calibration on its own. No further steps are required.
      • If, before the calibration ends, you make any other change to the SharpV configuration, the calibration is restarted.