Configuring SharpZ3 analytics in the SharpOS 13.7 Sharp Portal - SharpZ3 | SharpOS 13.7

AutoVu™ SharpZ3 Administrator Guide 13.7

Product
SharpZ3
Content type
Guides > Administrator guides
Version
13.7
Language
English
Last updated
2024-09-06

To ensure the highest license plate read accuracy, you must let the SharpZ3 system know what type of license plates to expect in the image. If required, you can also configure the system to extract additional information from the image, such as the license plate origin and the vehicle type.

What you should know

AutoVu MLC™ (Machine Learning Core), which performs video analytics on SharpZ3 cameras, is retrained on a regular basis. As SharpZ3 OS updates are released over time, you will notice improvements in the effectiveness of certain analytics.

Procedure

  1. Log on to the Sharp Portal.
  2. From the Configuration menu, click Analytics > ALPR settings.
  3. From the Context menu, select which license plates will be most commonly read by the SharpZ3.
    NOTE: For more information on the available regional contexts, see Supported ALPR contexts.
    Tip: To help us improve the performance of regional contexts, click Configuration > Connectivity > Product improvement and register this camera to participate in the Product improvement program .
  4. From the Reading mode menu, select one of the following reading modes:
    Continuous
    Select this to capture plate reads continuously. This is the default setting.
    Conditional
    Select this to capture plate reads continuously as long as the selected input signal meets the defined high or low condition.
    Multiconditional
    Using this option, you can configure the system to generate reads from a specific camera only while an input's status meets the defined high or low condition. This option is required if using the AutoVu™ car camera switch.
  5. From the Read strategy list, select a read strategy:
    Slow moving vehicle
    Applicable for use in typical city, law enforcement, or mobile parking enforcement.
    Fast-moving vehicle
    Applicable for installations where vehicles travel at moderate to high speeds. For example, use this read strategy for law enforcement installations where vehicles might be traveling at high speeds in opposite directions.
  6. Under Analytics, select the contents of the plate you would like the SharpZ3 to attempt to read. You can select the following:
    NOTE:
    • For a full list of analytics and recommendations, see Supported ALPR analytics.
    • You can add the state, vehicle make, and confidence score as annotation fields in Security Center to query for this information in Security Desk reports.
    State
    Select this option if you want the SharpZ3 unit to attempt read the license plate origin. Depending on the region, this can refer to the issuing state, province, or country.
    NOTE: For more information, see Supported plate origin recognition.
    Vehicle make and model
    Select this option if you want the SharpZ3 unit to attempt to read the vehicle’s make and model, for example, Honda Civic.
    NOTE: This setting is not recommended for MLPI installations.
    Confidence score
    The SharpZ3 assigns a confidence score percentage to each license plate read. This value indicates how confident the SharpZ3 is in the accuracy of the read.
    NOTE: License plates that contain similar characters such as 8 and B are more difficult to read and generally produce reads with a lower confidence score.
    Vehicle type, color, and orientation
    Select this option if you want to include the vehicle type, color, and orientation with plate reads. Based on the context image, Sharp analytics can detect if the vehicle is, for example, a light yellow bus, or a dark green passenger vehicle. In addition, Sharp analytics can detect the vehicle orientation, for example, back, or front side.
    Vehicle accessories
    Select this option if you want license plate reads to indicate the presence of items such as hitches, spare tires, and other vehicle accessories.
    NOTE:
    • To enable this feature, contact your AutoVu representative.
    • Hitches must be at least 20 px high in the image for detection. Note that hitch detection can be affected by environmental factors such as snow and ambient lighting.
    • Using the vehicle accessories feature increases CPU usage and can affect the frame rate.
    The ALPR Settings page with the analytics options selected.
  7. Click Save.