SIRAM: Vehicle Recognition Software (OCR)
Within the SIRAM family of products there are different software modules that are integrated into
different Suites, but the main one is the OCR Software.
This is the software that allows for the automatic recognition of number plates for each capture unit or camera, it being possible to install multiple OCRs on the same computer or on different computers (Modular Architecture). The maximum number of OCRs per computer will depend on its specifications.
After the number plate is identified, this data can be sent to different types of software:
The Siram Access Control Suite (Suite v3 or Suite v4)
The Car Park Management software
The Security Access Points Control software
Other software (scales, video supervision, etc.)
HIGH TOLERANCE TO SLOPES (ANGLES)
The horizontal and vertical angle from the number plate is very important, because the larger this is, the more deformed the character will appear and the more complicated its recognition will be. Siram OCR makes automatic corrections that allow a high rate of recognition in situations with a large angle.
LARGE LANE WIDTH.
The lane width supported depends on the resolution of the camera and lens model used. Usually a lane width of 3.5 metres is recommended, but with a Mpixel camera and a varifocal lens the OCR can recognise the number plate in lane widths of up to 5 metres.
The lane width and distance to the point of capture determine the angle of incidence of the camera with the number plate, the maximum recommended angle being 30 degrees.
The system supports speeds up to 50 km/h, although it should be remembered that the vehicle speed determines the positioning of the capture unit in order to have enough frames to analyse.
Therefore, in an access situation with a barrier where the capture unit is 5m from the trigger point,
speeds over 20 km/h are not recommended. For an access point without a barrier and with speeds
of over 20 km/h, it is recommended to position the capture unit on a pole and at a greater distance
from the capture point (10 metres).
From the OCR interface it is possible to:
– See if the adjustment and zoom are correct.
– Adjust internal OCR parameters.
– Adjust Gain and Shutter of the camera.
– Perform a reset of the camera to the parameters.
– Perform a TEST of the OCR.
In this working mode, the OCR waits for a trigger or request to read the number plate in a certain number of images and send the final result. This final result is obtained by calculating the average of the OCR results in each of the individual images.
With the PRE and POST configuration parameters we can configure the number of images to analyse before and after the trigger. For example, with PRE=1 and POST=1, one image before the trigger and one image after the trigger will be analysed, as shown in the following diagram.
Through the application of the OCR we can observe the results of each image analysed and in this way appropriately configure the number of PRE and POST images.
In this working mode the OCR is performing readings of number plates on the images received continuously. The readings of each image are stored in an input buffer and then sent to a filtering stage. If the reading exceeds the filtering stage it is stored in an output buffer.
In PUSH mode, each time a number plate is detected and the filtering stage is exceeded, the result is sent. Therefore, PUSH mode does not require a request for the sending of the result.
CONTINUOUS PULL (CONTINUOUS AND TRIGGER COMBINATION)
Every time a number plate is detected and the filtering stage is exceeded, the Reading is stored in the timeline (buffer). When the trigger or request is received by the client, SIRAM OCR sends the result with the highest Reliability Rate of those stored in the Timeline.
Do you want to know more?
Give us your details and we will call you: