This document provides general and technical information required to install, interface and operate the See/Car Vehicle License Plate Identification DLL (Dynamic-Link Library) product.

2.References

2.1 See/Lane - Hi-Tech Solution’s Access control system

2.2 See/Park - Hi-Tech Solution’s parking management system

Note: In this document, all references to “Vendor” means Hi-Tech Solutions Ltd.

3.Overview

3.1.Product description

SEE/CAR is a License Plate Recognition (LPR) software/system that is used to automatically read a license plate on a car, as shown in the following illustration. It is available in several configurations. This manual refers to the Windows DLL product configuration.

3.2.System Architecture

The See/Car DLL (Dynamic-Link Library) is a subsystem in a user’s application. It runs in a PC under the User’s application program. It contains plate recognition software, with public function calls that the user can call from the application program.

The following illustration shows this configuration in a typical live system, where the User’s Application runs in a PC. It interfaces one or more cameras, captures the image containing the License Plate with a frame grabber board into memory. Then it calls See/Car DLL. The DLL processes the image and returns the recognition result and status.

Figure 3.2.1 : See Car DLL integration - On-Line Mode

Another option (off-line mode) is to processes image files previously captured in the field. The User’s program opens each image file, reads the pixels into memory and calls the DLL.

Figure 3.2.2 : See Car DLL integration - Off-Line Mode

3.3.Sample Applications

The system can be part of the following applications (a sample of possible use of the system using See/Car DLL):

3.3.1.Automated Parking Lot

The cars reaching the entrance gate will trigger the projector light, which will turn on. The camera will ‘see’ the plate and the User’s program will capture an image. It will call the DLL to identify its number, and send the recognition string via RS232 or network to the host computer. The host computer will open the gate. When the car exits through the exit gate, the same process is repeated. The host computer in this case will locate the car in the stored numbers, and will calculate the parking fee and open the gate (or in case of membership parking, will just open the gate).

The history of car entries and exits will be used for parking fee calculations, data logging, marketing information and statistics. Additionally, the list can be compared at any time with actual parked cars for security reasons.

In this configuration, it is assumed that a host computer that receives the car plate readings and performs the required financial calculations, and also controls the gates controls the parking lot. The PC using the See/Car DLL works as the ‘eyes’ for this host computer.

Figure 3.3: Gate Access control illustration

Note that this application is performed by our access control application, See/Lane (ref #2.1). This application handles the DLL and the hardware and operates as a stand-alone access control system, which can be integrated into different types of applications. Another product is See/Park (ref #2.2) which has additional parking management functions. Please refer to the applications’ documentations.

3.3.2.Gate to Factory or secured area

An alternative operation mode can be that the PC with the See/Car DLL works as a stand-alone system, without a host computer. In such applications, a pre-stored list of authorized cars will be stored in the PC as a file. The car will reach the gate, turn the projector on, and will be identified by the DLL. Then, the PC will open the gate and store or print the identified vehicle.

This application is suitable for army installations, factories and other secured facilities.

Note that this application is also supported by See/Lane application.

3.3.3.Toll road

The system can be used to data collect the car plates that pass a toll road gate, for billing or data collection.

3.3.4.Police uses and Law Enforcement

3.3.4.1.Stolen Cars

A predefined list of stolen cars can be entered to the system, and its detection will signal an alarm when such a car will pass.

3.3.4.2. Speed Violations

Another system application can be used for law enforcement, either for speed violation or entry to illegal traveling zones. In this case the system can identify cars on-line, or process image files off-line (for ticketing).

The automatic identifications accelerate the processing of the violations. The manual operation takes a few minutes per image, while the automatic processing can be performed at a rate of several images per second. This is a 1:1000 quicker turnaround rate.

3.4.See/Car/DLL advantages

The See/Car/DLL has the following advantages over existing LPR software:

· simple integration into the existing or new applications

· has a high recognition rate (multiple checks for each letter/digit)

· covers a wide range of plate size (80-300)

· deals with very small plates (down to ~80 pixels per plate on typical plates)

· deals with wide range of contrast images

· allowable rotation ± 30 degrees (depends on plate type)

· configurable operation (using standard file)

· fast response - one of the fastest packages available. Pentium 700Mhz at typical 60msec per image to return the recognition string. This allows the user’s application to utilize most of the PC resources.

· low cost solution - standard PC with minimal resources

· Adaptable to different country standards (with a simple configuration file change)

· Available in 32bit versions, for use with standard compilers

· source code for sample applications is available (cuts time-to-market)

· Highly integratable applications (see See/Lane)

4.Architecture

This section describes the architecture of the See/Car DLL.

4.1. Overview

4.1.1.Functions

The system performs the following functions within the DLL:

· Image handling - handles the input image and prepares it for the processing

· Image Enhancement - improves the quality of the image

· Plate finding - locates the license plate within the image

· Slope calculation - measures the angle of the plate

· Characters detection - locates each of the plate’s characters

· Characters identification - identifies each of the detected characters

· Validation - compares the results

· Output - sends the results back to the User’s program

The effect of this operation is like a reduction program: a huge array of information (the image) is converted to a short ASCII string containing the identified plate. The conversion ratio is roughly 512 X 512 (average image size in pixels, each pixel a Byte) to about 10 characters!

4.1.2.Application program

Note that the DLL does not capture the image (via a frame grabber for on-line applications or via a file read for off-line applications). This is the role of the application program which performs all the hardware and system functions of the application. Furthermore, the DLL does not process the recognition results and outputs it to the display or computer network.

Sample applications in source file will be supplied to the developer as part of the DLL package. Other applications, such as a parking lot system application, can be licensed.

4.1.3.Structure

The DLL is composed of the following elements:

· DLL object code which performs its functions

· Security plug which is required to be plugged in the parallel port

· “Format.ini” configuration file - which describes the parameters of the plate and identification. A detailed explanation is provided in the following subparagraph.

4.2.Format.ini Configuration File

Since the DLL is designed to operate in various configurations, installations and local standards, a special text file is available to optionally change the configuration and installation of the system. The file is in a standard initialization (“.ini”) file format as commonly used in Windows environment. This file is a text file with a set of fields (each enclosed in square[ ] brackets) and a set of assignments (par _I= value _j). Comments are enclosed after “;”. The file can be edited using standard text editors.

Thus, the vendor or user can shape the characteristics of the system in order to fit it to the required environment.

This file (format.ini) has the following types of parameters:

· Formats - Define formats that describe the local standard. (For example, in a certain country there are 3 different standards of the plate then each standard is described as a separate format). This is prepared by the Vendor before shipping the product, and tested against a set of sample test images.

· Internal Parameters - a set of values used for the optimized execution of the recognition software. This also is prepared by the vendor and optimized for a certain country standard.

· External Parameters - a few parameters are site specific and can be either determined by the Vendor, or set on site by the installer. These are described in a later section.

A short description of some of the parameters are provided in the next section.

4.3.DLL Input and Output

The DLL’s main function (sc_get_id_string) accepts the following inputs and outputs, as described in seecar.h (the header file for the DLL):

SEECAR_RET_STATUS SEEAPI sc_get_id_string (

byte MEM* p_byte, /* source image bytes buffer */

id_string plate_string, /* resulting plate string */

byte* plate_type /* resulting plate type */

);

The function receives:

(1) a pointer to the input image in memory (*p_byte),

(2) a pointer to the returned string (plate_string),

(3) The identified format is returned with plate_type pointer (one of the possible formats, as described in the formats.ini file). This information is usually not used by the User’s application.

In addition to the parameters, a return code (SEECAR_RET_STATUS) is also returned with the status of the recognition. The possible error codes are:

0=SC_SUCCESS, /* successfuly proccessed */

1=SC_FAILURE, /* failed to process */

2=SC_NO_PLUG, /* security plug not found */

3=SC_NO_MEM /* not enough memory */

4.4.Image type

The major parameter that is passed to the DLL is a pointer to the image stored in memory. The contents of the image were previously captured or read by the calling application. The image should have the following characteristics:

· 256 levels (8-bit) gray-level, where: white is 255, black is 0.

· Image size is described in format.ini configuration file (e.g., 512 X 512)

· both frame (both even and odd lines) or field (only even or odd lines) type images could be passed to the DLL (a parameter in the format.ini configuration file)

In case of a B&W camera, the images are directly extracted to the image memory. In case of color images, the application program should convert the color information into gray-level (using the palette information and using the intensity equation conversion).

4.5.Other functions

Refer to SeeCar.h include file for a updated and formal description of the interface with the DLL. This header file includes additional function calls that are available in the DLL.

5.Changing Site-specific Parameters

This section describes how certain parameters should be changed in specific sites. Note that in most sites the values can be the same. Furthermore, the Vendor can make this changes for you on supplied sample images. Always keep a history of previous versions in order to minimize installation problems.

5.1.Captured image size

The following lines describe the captured image size in pixels:

capture_width = 768 ; captured picture size

capture_height = 288 ;

Note that this example if for a standard CCIR field image. (The field data is an image with even or odd lines only).

5.2.Full-Frame or Field

The DLL can accept either Full-Frame images (frame=1) or single Field images (Frame=0). The line that requires this setting is:

frame = 0 ; whole frame = 1; even/odd field only = 0

5.3.Deinterlace

In case of full-frame only, you can determine whether to use the whole frame image contents (even and odd rows) or to perform deinterlacing correction. In the latter case the parameter should be set to 1. This prevents the effect of the moving car, since there is a time delay between even and odd rows in interlaced cameras.

deinterlace = 1

For progressive-scan or Non-interlaced cameras, the value could be 0 (no deinterlacing).

5.4.Margins

The image margins should be defined in order to prevent the DLL to process titles or data that is known to exist on all images. For example, a law-enforcement image contains the speed violation data on the bottom line (20 pixels). In this case the margin in the bottom side should be set to 20:

bottom_margin = 20 ;;due to bottom speed-data text

For all 4 sides of the image, this example could be expanded as follows. Note that the value 3 is the minimum margin.

left_margin = 3 ;;the minimum margin

right_margin = 14

top_margin = 3

bottom_margin = 20

5.5.Max/Min Plate Length & Height

In order to prevent the software to lock on false alarms, the following statements could limit the size of the plate (in pixels):

min_hor_plate_length = 70

max_hor_plate_length = 200

min_hor_plate_height = 20

max_hor_plate_height = 60

6.

See/Lane application is also available in a simulated version - which is available on our web-site.

Visit our Home page http://www.I-Cube.co.za and download recognition demo players, see sample applications and get updated information. Qualified customers also receive on-line updates in the support page.

Visit our Home page http://www.I-Cube.co.za and download:

demo applications

See/Lane simulation program

Latest tools and releases (in our on-line support page)

We also recommend to install pcAnywhere with TCPIP connection. This software package will allow us to view the on-line operation, fine-tune the parameters, download images and update the software. See more details in the web site support page.