Hikvision of the ANPR camera

2019: The presentation of ANPR cameras in the Russian market

On July 15, 2019 the Hikvision company presented in the Russian market channel ANPR cameras for small and average projects. Devices have the built-in mechanism of recognition of car numbers (ANPR) which is based based on algoritmovgluboky training (Deep Learning).

As noted in Hikvision, Deep Learning technologies allow not only to read out registration number of the vehicle, but also to reveal the CU without number, to define the direction of the movement. Also algorithms of deep learning are capable to distinguish cars and motorcycles.

According to the producer, intelligent ANPR cameras can also be used to monitoring of traffic, control of observance of traffic regulations, fixing of violations with the broad range of scenarios, such as paid and city roads, the mixed types of roads, tunnels, parkings, etc. The corporate Hikvision technology – DarkFighter – provides higher sensitivity of a matrix (0.001 lx) that allows ANPR cameras to transfer the qualitative and detailed image even in low-light-level conditions.

Features of ANPR cameras mentioned by the producer:

• High resolution: 2 Megapixels
• The motorized lens with an auto focus. Two versions of cameras for different scenarios of use:
  • the lens of 2.8-12 mm – is suitable for the solution of narrowly targeted tasks (control of input and output zones, points of collection of a payment, gas stations, etc.)
  • the lens of 8-32 mm – is suitable for the solution of a wide range of tasks for ensuring transport security (control of city roads, toll highways, roads of the mixed type, etc.)

• The codec of compression H.265 – transfer of the high-quality image
• High frequency of a frame is up to 50 Hz: 1920×1080 @ 50fps allows to receive the image with quickly moving CU without delay
• Function of three-dimensional noise reduction 3D DNR, stabilization of the image of EIS and broad dynamic range of WDR 140 dB provide high quality of video filming
• IK-illumination up to 50 m (2.8-12 mm) / 100 m (8-32 mm)
• The aperture of F1.2 provides broader disclosure of a lens for the qualitative image even in low-light-level conditions
• Algorithms of deep learning will provide the high accuracy of capture (> 99%) and recognitions of numbers of cars (> 98.5%)
• Smart function: determination of the CU without car number, direction finding of the movement of the car, determination of the motorcycle
• Wiegand interface
• Open access to SDK and ISAPI provides ample opportunities for integration
• Protection of classes IP67 and IK10 allows the device to work in any climatic conditions
• Support of PoE simplifies installation process
• Broad range of working temperatures: from −40 °C up to +70 °C.