Home

Manufacturer

ELVEES

124460, Moscow, p/b 19 Moscow, Zelenograd, Southern industrial zone, Technopark "Zelenograd", p.4922, b. 2

elvees@made-in-zelenograd.com

Senesys-Avto is Senesys access control system integrated with ĎAuto Numberí system. Senesys access control system in Senesys Avto is used for creating and keeping a single customer and cars data base, for differentiate access levels across the territory zones and time, for managing the execution units (gateways, lifting gates), keeping the detailed log of the events and flexible reporting to the security service. ĎAuto-Numberí being the part of the Senesys-Avto provides reliable recognition of the license plate numbers of the vehicles entering or leaving the territory.

Principle of operation

The surveillance camera, mounted at the control post captures the approaching vehicle. Video data are loaded to ĎAvto Numberí server where the received data are preprocessed and the license plate is analyzed. The recognized vehicle license plate number as well as license plate type and corresponding channel identificator data are transmitted to Senesys server.

Senesys system checks whether the vehicle is in the data base. If the vehicle is registered and is authorized to enter the system commands to open the lifting gate. Otherwise the operator is offered to add the vehicle in the data base, open the execution unit manually or reject the access. The passage as well as all operatorís actions are logged.

Appearance verification module, plugged-on Senesys system, helps to avoid license plate fraud. The picture of the vehicle from the video camera is compared with the photo of the vehicle which the recognized license plate belongs to according to the data base. In case of a mismatch the operator is alerted.

General capabilities of the system

  • Client and vehicle data base keeping with detailed information about the vehicle (car make, model, color, etc.) and people allowed to driving
  • Vehicle identification by license plate number and/ or proximity card, additional verification by appearance
  • Automatic access to the territory provided the identified vehicle has the necessary authorities. Manual control of execution units.
  • Vehicle passage logging including freeze frames, date, time, access reason (operatorís instruction, for example)
  • Vehicle presence control on the guarded territory and parking time accounting

License plate number recognition technology

The algorithm of license plate number recognition is Elvees in-house development. The main competitive difference of the development is a unique grayscale image character analysis technology that considerably improves recognition quality.

Binary methods often lead to character rupture and patching as well as symbol 'sticking' to the border of the license plate that results in misrecognition of the license plate number. Grayscale image recognition technology developed by Elvees specialists excludes data loss due to wrong binarization threshold choice and operates with dirty, weak-contrasting or poorly illuminated license plates successfully. The recognition algorithm operates properly at large viewing angles and in bad weather conditions with natural noise, for example, rain, snow or fog.

License plate recognition is carried out in several steps.

Step 1 ó Selection of supposed license plate areas

Areas presumably containing license plate are marked on the frames received from the video camera. The marked areas may also contain other image elements having regular structure, for, example, radiator guard or inscription on the vehicles.

Step 2 ó License plate type identification

Each area marked at the previous step is analyzed for coincidence with any template containing license plate type description. The template specifies the size, structure, character and background color of the license plate, demarcation strip position, fastener apertures, etc. Areas not matching any possible type are excluded from subsequent examination. Exact license plate width and height, slope and inclination angles as well as possible bending are determined. License plate character positions are identified.

Step 3 ó Character recognition

The next step is separate character recognition taking into account that each character position may contain character of a certain type; some positions may contain only numerals, the others may contain certain letters only. Validity coefficient is calculated for each recognized character.

Step 4 ó Final number formation

License plate number is recognized in each frame but the final result is formed after integrated analysis of images referring to the vehicle passage. In some frames the license plate may be overexposed by the headlights, in other a spot of light or a shadow prevented one or more characters from recognition. Joining the results from multiple frames improves processing quality and issues a trustworthy recognized number for the vehicle as well as determines its driving direction.

Send enquiry

Made in Zelenograd

free counters