KR102684524B1 - Supersaturation removal method for Vehicle number plate and, vehicle number recognizer therewith - Google Patents

Abstract

The present invention controls the polarizing lens installed in front of the lighting lens and the polarizing filter installed in front of the camera lens, effectively eliminating the oversaturation and spot phenomenon that occurs in vehicle license plates, thereby improving the accuracy and reliability of license plate recognition. This relates to a method for removing license plate oversaturation that can significantly reduce the false recognition rate while increasing the number plate recognition rate, and a license plate recognition system using the same.

Description

Supersaturation removal method for Vehicle number plate and, vehicle number recognizer therewith}

The present invention relates to a method for removing license plate oversaturation and a vehicle license plate recognition device using the same. In detail, the polarizing lens installed in front of the lighting lens and the polarizing filter installed in front of the camera lens are controlled to prevent oversaturation generated in the vehicle license plate. It relates to a license plate oversaturation removal method that can significantly reduce the false recognition rate while increasing the accuracy and reliability of license plate recognition by effectively removing the oversaturation and spot phenomenon, and a license plate recognition device using the same.

Recently, as image processing technology has developed and the camera industry has developed, cameras are being widely installed in various places and spaces for various purposes.

In particular, license plate recognition devices are installed by default at the entrances and exits of parking lots, and since these license plate recognition devices are used as a basic medium for identifying vehicles entering and leaving the parking lot, the demand for them is increasing exponentially.

Such license plate recognition devices typically consist of a lighting device that emits light and a camera that acquires a light image reflected by the lighting device. As a result, the light is intensively reflected on the vehicle license plate that is the subject of photography, causing oversaturation of lighting and A spot phenomenon occurs, causing a problem in which the number recognition rate decreases.

Meanwhile, polarizing filters (PL filters, polarized light filters) are widely used in camera modules due to their advantage of being able to obtain clear and dark images by blocking light reflected from the surface of objects such as roads or windows. It is mainly configured to be detachable and rotatable, but this conventional method is cumbersome and complicated to use because the user must manually rotate the direction of the polarizing filter, and it is difficult for users who are not skilled in photography technology to use. This happens.

Figure 1 is a block diagram showing a license plate recognition camera system disclosed in Domestic Patent No. 10-2190544 (title of the invention: license plate recognition camera system).

The license plate recognition camera system (hereinafter referred to as the prior art) 100 of FIG. 1 includes an optical lens 105, a filter switching unit consisting of a blocking filter 111 that blocks infrared or visible light passing through the optical lens, and a polarizing filter. One or more cameras 120 that have a lens unit (L) with a built-in lens unit (110) installed to photograph and transmit a license plate, and an image sensor 130 that detects an image captured from the lens unit (L), It is equipped with an image processing unit that receives the digital signal converted from the image sensor 130 and processes the image, and an image collection unit that receives the processed image signal, compresses the image, and stores the compressed image, and transmits the image information to an external image display device. It consists of an image processing means 140 formed with an image output port for output.

The filter switching unit 110 includes an infrared blocking filter, a visible light blocking filter, and a polarizing filter. During the day, it transmits visible light and blocks infrared and reflected light to generate a color image, and at night, it blocks visible light and creates an infrared light. It transmits and creates a black and white image.

The prior art (100) configured in this way can clearly read vehicle license plates by capturing and acquiring clear images from surveillance cameras, etc., even in the presence of fine dust, heavy rain, fog, reflection of vehicle lights, and miscellaneous light, regardless of day or night. It has the advantage of being able to receive image information and process images even for vehicles whose license plates are not in constant positions.

However, the prior art (100) only describes a configuration that simply includes an infrared cut-off filter, a visible light cut-off filter, and a polarizing filter that are usually installed in a camera, and is a technology on how to remove the lighting oversaturation and spot phenomenon of license plates. Since this is not described at all, it has a structural limitation in that the number recognition rate is lowered due to oversaturation and spots in the license plate image.

In particular, when the car license plate is reflected by lighting at a certain distance from the camera at the entrance to the parking lot, problems with oversaturation and spotting become more severe.

The present invention is intended to solve this problem, and the problem of the present invention is that the polarizing filter installed in front of the camera lens is configured to filter the vertical vibration wave component of the input optical image, thereby preventing oversaturation and spot phenomenon in the license plate image. The purpose is to provide a method for removing license plate oversaturation that can significantly increase the accuracy and reliability of license plate recognition by effectively removing the license plate oversaturation, and a license plate recognition device using the same.

In addition, another problem of the present invention is that the polarizing lens installed in front of the LEDs is configured to remove the vertical vibration wave component of the optical signal emitted from the LEDs, but to pass the horizontal vibration wave component, thereby further improving the license plate recognition rate. The purpose is to provide a method for removing license plate oversaturation and a license plate recognition device using the same.

In addition, another problem to be solved by the present invention is a method of removing license plate oversaturation that can further increase the license plate recognition rate by effectively blocking sunlight by having the camera unit further include a solar filter installed in front of the polarizing filter, and a license plate recognition device using the same. It is intended to provide.

In addition, another problem solved by the present invention is that the polarizing filter of the camera part is configured to be rotatable, and at the same time, the controller is configured to calculate the optimal rotation angle of the polarizing filter at each preset period (T), thereby maximizing the number recognition rate compared to the same environment. The purpose is to provide a method for removing oversaturation of license plates and a license plate recognition device using the same.

The solution of the present invention for solving the above problem is a license plate recognizer for recognizing the license plate number of a detected vehicle: the license plate recognizer includes a camera that acquires an optical image by photographing the detected vehicle, and a lens of the camera. A camera unit installed in the front, including a polarizing filter that filters vertical/horizontal vibration wave components of an input optical image, and a filter rotation control unit that controls rotation of the polarizing filter; A controller that recognizes the license plate number of the detected vehicle by analyzing the optical image obtained by shooting the camera unit; It is executed under the control of the controller, and is installed in front of the LEDs that emit optical signals and the LEDs, removes the vertical/horizontal vibration wave components of the emitted optical signals, and passes the horizontal vibration wave components of the optical signals. It includes a lighting unit including a polarizing lens, and the controller includes a shooting control unit that controls the operation of the camera unit; a lighting control unit that controls the operation of the lighting unit; an image analysis unit that analyzes the optical image passing through the polarization filter of the camera unit; a license plate extraction unit that extracts a license plate image from an optical image with reference to the data analyzed by the image analysis unit; A license plate recognition unit that recognizes the license plate number by analyzing the license plate image extracted by the license plate extraction unit; The operation method (S38) of the rotation angle optimization unit, which includes a rotation angle optimization unit that is executed every preset period (T) and is performed by the controller, includes initializing the rotation angle (θ) of the polarization filter to 0° in step 381. (S381); Step 382 (S382) of controlling the filter rotation control unit to rotate the polarization filter at the rotation angle (θ = 0°) initialized in step 381 (S381); When the polarization filter is rotated in step 382 (S382), controlling the camera unit to capture a license plate (S383); When photography is performed by the camera unit through step 383 (S383), step 384 (S384) of extracting a license plate image by analyzing the obtained optical image; Step 385 (S385) of detecting the light saturation of the license plate image extracted by step 384 (S384); Step 386 (S386) of determining whether the current rotation angle (θ) of the polarization filter is '90°'; In step 386 (S386), the process proceeds when it is determined that the current rotation angle (θ) of the polarization filter is not '90°', and a preset increase value ( Step 387 calculates the next rotation angle (θ', θ' = θ + α) of the polarization filter 75 by adding α), then returns to step 382 (S382) and repeats the subsequent process. (S387); In step 386 (S386), when the current rotation angle (θ) of the polarization filter is determined to be '90°', light-saturation values for each rotation angle of the polarization filter are analyzed, and among these, light- After calculating the rotation angle (θ) with the lowest degree of saturation, step 387 (S387) of determining the calculated rotation angle (θ) as the optimal rotation angle of the polarization filter in the next cycle, wherein the filter rotation control unit The optimal rotation angle value is not fixed but is changed by rotating the polarizing filter according to the optimal rotation angle value determined by the rotation angle optimization unit.
In addition, in the present invention, the lighting unit preferably further includes a horizontal diffusion lens (CLA, Collimating Lens Array) that diffuses the horizontal component of the optical signal emitted from the LEDs.

delete

delete

In addition, in the present invention, it is preferable that the camera unit further includes a solar filter that is installed in front of the polarizing filter and filters solar light.

Additionally, in the present invention, the LEDs are preferably IR LEDs (Infrared LEDs).

delete

Additionally, in the present invention, the rotation angle of the polarizing filter preferably ranges from 0° to 90°.

According to the present invention, which has the above problems and solutions, the polarizing filter installed in front of the camera lens is configured to filter the vertical vibration wave component of the input optical image, thereby effectively eliminating the oversaturation and spot phenomenon of the license plate image, thereby The accuracy and reliability of number recognition can be significantly improved.

In addition, according to the present invention, the polarizing lens installed in front of the LEDs is configured to remove the vertical vibration wave component of the optical signal emitted from the LEDs but to pass the horizontal vibration wave component, thereby further improving the license plate recognition rate.

In addition, according to the present invention, the camera unit further includes a solar filter installed in front of the polarizing filter, thereby effectively blocking sunlight, thereby further increasing the license plate recognition rate.

In addition, according to the present invention, the polarizing filter of the camera unit is configured to be rotatable, and the controller is configured to calculate the optimal rotation angle of the polarizing filter at each preset period (T), thereby maximizing the number recognition rate compared to the same environment.

Figure 1 is a block diagram showing a license plate recognition camera system disclosed in Domestic Patent No. 10-2190544 (title of the invention: license plate recognition camera system).
Figure 2 is a configuration diagram showing a license plate recognition device according to an embodiment of the present invention.
Figure 3 is an example diagram showing the license plate recognition device of Figure 2 installed.
Figure 4 is an exemplary diagram showing the oversaturation phenomenon of a conventional vehicle license plate.
FIG. 5 is an example diagram illustrating a process in which the vertical component of an IR optical signal is filtered by the polarizing lens of FIG. 2.
FIG. 6 is an example diagram for explaining the process of generating a vertically polarized image by the polarizing filter of FIG. 2.
Figure 7 (a) is an example diagram showing an optical image when the solar filter of Figure 2 is not applied, and (b) is an exemplary diagram showing an optical image when the solar filter is applied.
FIG. 8 is a graph for explaining the characteristics of the solar filter of FIG. 2.
Figure 9 (a) is an example showing a license plate image with oversaturation when the present invention is not applied, and (b) is an example showing a license plate image with the oversaturation phenomenon removed when the present invention is applied. It's a degree.
FIG. 10 is a block diagram showing the controller of FIG. 2.
Figure 11 is a flow chart showing the operation process of the rotation angle optimization unit of Figure 10.

Hereinafter, an embodiment of the present invention will be described with reference to the attached drawings.

Figure 2 is a configuration diagram showing a license plate recognition device according to an embodiment of the present invention, Figure 3 is an exemplary diagram showing the license plate recognition device of Fig. 2 installed, and Figure 4 is an exemplary diagram showing a conventional license plate oversaturation phenomenon. .

The license plate recognition device (1), which is an embodiment of the present invention, is installed on the entrance and exit roads of a parking lot, and is a device for recognizing the license plate number by photographing the front or rear of a vehicle passing a preset position of the entrance/exit road. , In detail, the polarizing lens 53 installed in front of the lighting lens and the polarizing filter 75 installed in front of the camera lens 71 are controlled to prevent oversaturation and spot phenomenon occurring in license plates. This is to increase the accuracy and reliability of license plate recognition by effectively removing the license plate number and at the same time significantly reduce the misrecognition rate.

In addition, as shown in Figures 2 and 3, the license plate recognition device (1) of the present invention is installed at the parking lot entrance and transmits an IR optical signal to the front of the entering vehicle or the rear of the exiting vehicle under the control of the controller (3), which will be described later. The lighting unit 5 emits an optical signal of the horizontal vibration wave component by filtering the vertical vibration wave component of the emitted optical signal, and acquires an image by photographing the front of the entering vehicle or the rear of the exiting vehicle. The camera unit 7 acquires a horizontal polarized image by filtering the vertical vibration wave component of the camera unit 7, and analyzes the horizontal polarized image obtained by shooting and filtering the camera unit 7 to determine the license plate number of the entering or exiting vehicle. After extraction, it is composed of a controller (3) that analyzes the extracted license plate number to recognize the license plate number and simultaneously manages and controls the operation of the lighting unit (5).

FIG. 5 is an example diagram illustrating a process in which the vertical component of an IR optical signal is filtered by the polarizing lens of FIG. 2.

The lighting unit 5 of FIGS. 2 and 5 includes IR LEDs (Infrared LEDs) 51 that emit IR optical signals toward the front, that is, toward the entering/exiting vehicles, and are disposed in front of the IR LEDs 51. , It consists of a polarizing lens 53 that filters the vertical vibration wave (vertical direction) component of the optical signal L emitted from the IR LEDs 51.

At this time, as shown in FIG. 5, the polarizing lens 53 filters the vertical vibration wave component of the IR optical signal emitted from the IR LEDs 51 of the lighting unit 5, so that the horizontal vibration wave component of the IR optical signal is transmitted to the vehicle. It will be displayed as a license plate.

When the lighting unit 5 emits light in a strobe manner, oversaturation and spotting may occur in the captured license plate image, and this oversaturation and spotting phenomenon in the license plate image acts as a major cause of license plate unrecognition. do.

Generally, in the case of low-illuminance conditions such as at night, lighting is emitted from the lighting unit 5, and the lighting light emitted from the lighting unit 5 depends on various variables such as the direction of the body of the subject (vehicle) and the shape and structure of the front vehicle body. reflected in various directions.

However, since the license plate P is formed of a square plate with a length (horizontal) greater than the width (vertical), the lighting unit 5 of the present invention installs a polarizing lens 53 in front of the IR LEDs 51, and polarizes the light. The lens 53 filters the vertical vibration wave component of the IR optical signal, so that the IR optical signal of the horizontal vibration wave component is emitted.

At this time, when the IR optical signal irradiated from the lighting unit 5, that is, the IR optical signal with a strong horizontal vibration wave component, is reflected by the vehicle body, the reflected light has the characteristic of increasing the vertical vibration wave component, and accordingly, In the present invention, a polarizing filter 75 is installed in front of the camera lens 73 of the camera unit 7, and the polarizing filter 75 is configured to filter the vertical vibration wave component of the reflected light, thereby preventing oversaturation and spot phenomenon of the license plate. It is possible to effectively prevent the phenomenon of vehicle license plate number not being recognized.

Meanwhile, the lighting unit 5 may further include a horizontal diffusion lens (CLA, Collimating Lens Array) for diffusing the emitted IR optical signal in the horizontal direction, and the horizontal diffusion lens (CLA) is wider than the vertical length. Considering the characteristics of the license plate, which has a large horizontal length, the light distribution for the license plate can be increased by spreading the IR optical signal in the horizontal direction, but it has the disadvantage of high light loss.

FIG. 6 is an example diagram for explaining the process of generating a vertically polarized image by the polarizing filter of FIG. 2.

The camera unit 7 of FIGS. 2 and 6 includes a camera 71 that photographs the front or rear of a vehicle entering a preset position, a camera lens 73 installed in front of the camera 71, and a camera lens ( A polarization filter (75) installed in front of the camera (73), filters the vertical vibration wave component of the light image of the subject to create a horizontal polarization image, and then transmits the generated horizontal polarization image to the camera lens (73), It consists of a filter rotation control unit 77 that controls the rotation of the polarizing filter 75, and a solar filter 79 that is installed in front of the polarizing filter 75 and filters sunlight from the subject.

At this time, although not shown in FIGS. 2 and 6 for convenience of explanation, the camera unit 7 includes a CMOS image sensor (Complementary MOS image sensor) that converts the optical image passing through the camera lens 73 into an electrical digital image signal. (not shown), a PCB board (not shown) electrically connected to the CMOS image sensor, and an image signal processor (not shown) that performs color compensation, black label compensation, and color interpolation on the digital image signal converted by the CMOS image sensor. may further include, and additional means for processing the optical image are not limited thereto, and various known technologies and methods may be applied.

The polarization filter 75 filters the vertical vibration wave component of the subject that has passed through the solar filter 79, generates a horizontal polarization image, and then transmits the generated horizontal polarization image to the camera lens 73.

In other words, the IR optical signal emitted from the IR LEDs 51 of the lighting unit 5 passes through the polarizing lens 53, as shown in FIG. 6, and the horizontal component IR optical signals are irradiated, and these IR optical signals As the optical signal is reflected by the license plate (P), it is converted into reflected light with a high vertical vibration wave component, and the reflected light with a high vertical vibration wave component is filtered by the solar filter 79 and then filtered by the polarization filter 75. ), the vertical vibration wave component is filtered, and the horizontal vibration wave component is transmitted to the camera lens 73.

The filter rotation control unit 77 controls the rotation of the polarizing filter 75.

At this time, if the rotation angle of the polarization filter 75 is 0°, the vertical vibration wave component of the reflected light is filtered by the polarization filter 75, and if the rotation angle of the polarization filter 75 is 90°, the polarization filter 75 The horizontal oscillatory wave component of the reflected light is filtered.

This filter rotation control unit 77 is mechanically coupled to the polarizing filter 75 and rotates the polarizing filter 75 itself, or arranges the LCD liquid crystal in front of the polarizing filter 75, so that the LCD liquid crystal displays according to the application of voltage. Various known methods can be applied, such as a configuration that delays the phase of an optical image of a subject passing through liquid crystal.

For example, the filter rotation control unit 77 is a polarizing filter 75, as disclosed in Domestic Patent No. 10-2281517 (title of the invention: Camera module capable of automatic filter replacement) applied and registered by the present applicant. After installing the LCD liquid crystal in front of the LCD liquid crystal and electrically connecting the '+' and '-' electric lines to the LCD liquid crystal, it is configured to delay the phase of the optical image by 90° depending on whether a predetermined starting voltage (3V) is applied. This LCD liquid crystal can be used 1) when the horizontal vibration wave component of the input optical image is strong, without applying the starting voltage, and 2) when the vertical vibration wave component of the input optical image is strong, with the starting voltage applied. It can be configured to delay the phase of the optical image by 90°.

Figure 7 (a) is an example diagram showing an optical image when the solar filter of Figure 2 is not applied, (b) is an exemplary diagram showing an optical image when the solar filter is applied, and Figure 8 is an exemplary diagram showing an optical image when the solar filter of Figure 2 is not applied. This is a graph to explain the characteristics of the solar filter in Figure 2.

Looking at the solar filter 79 with reference to FIGS. 7 and 8, the solar filter 79 is disposed in front of the polarizing filter 75 and filters sunlight from the input light image, especially with a wavelength of 700 to 800 nm. By filtering the optical signal, it is possible to effectively remove oversaturation, spots, and smear phenomena in license plate images caused by diffuse reflection of sunlight.

Figure 9 (a) is an example showing a license plate image with oversaturation when the present invention is not applied, and (b) is an example showing a license plate image with the oversaturation phenomenon removed when the present invention is applied. It's a degree.

Referring to Figures 9 (a) and (b), in the present invention, the lighting unit 5 includes a polarizing lens 53, so that the horizontal vibration wave component of the emitted IR optical signal is formed to be high, and the camera unit (7) includes a polarizing filter 75, which filters the vertical vibration wave component of the reflected light, thereby dramatically reducing oversaturation and spotting of the license plate image.

FIG. 10 is a block diagram showing the controller of FIG. 2.

As shown in FIG. 10, the controller 3 includes a control unit 30, a memory 31, a data input/output unit 32, a shooting control unit 33, a lighting control unit 34, an image analysis unit 35, It consists of a license plate extraction unit (36), a license plate recognition unit (37), and a rotation angle optimization unit (38).

The control unit 30 is the operating system (OS) of the controller 3, and the control object (31), (32), (33), (34), (35), (36), (37), (38) Manage and control their operations.

Additionally, when the control unit 30 receives a polarized image from the camera unit 7 through the data input/output unit 32, it inputs the input polarized image to the image analysis unit 35.

Additionally, when the license plate number is recognized by the license plate recognition unit 37, the control unit 30 transmits the recognized license plate number to an external node.

In addition, the control unit 30 executes the rotation angle optimization unit 38 at each preset period (T), and uses the rotation angle optimal value set by the rotation angle optimization unit 38 as the filter rotation control unit of FIG. 2 (described above). 77).

The memory 31 stores the polarized image obtained by shooting with the camera unit 7.

Additionally, the memory 31 stores vehicle number information recognized by the number recognition unit 37.

Additionally, the optimal rotation angle value detected by the rotation angle optimization unit 38 is stored in the memory 31.

The data input/output unit 32 inputs and outputs data to the camera unit 7 and the lighting unit 5.

The shooting control unit 33 controls the operation of the camera unit 7, and the lighting control unit 34 controls the operation of the lighting unit 5.

For example, when an entering or exiting vehicle passes a preset location, the shooting control unit 33 outputs a trigger signal to the camera unit 7 to enable shooting, and the lighting control unit 34 ) can control the lighting unit 5 to emit light when an entering/exiting vehicle is detected if the illuminance value sensed by the illuminance sensor (not shown) is less than the threshold.

As another example, optical filters are generally effective in preventing oversaturation, but have the characteristic of darkening the image due to the light blocking effect. Therefore, the lighting control unit 34 controls the amount of light by controlling the aperture, etc. to increase the amount of light. It can be configured to do so.

When the image analysis unit 35 receives a polarized image that has passed through the polarization filter 75 of the camera unit 7, it analyzes the input polarized image.

In this case, analysis refers to image pre-processing work to facilitate the extraction and recognition of license plates in polarized images.

The license plate extraction unit 36 extracts the license plate from the polarized image by referring to the data analyzed by the image analysis unit 35.

The number recognition unit 37 uses a preset number recognition algorithm to analyze the license plate extracted by the license plate extraction unit 36 and recognizes the license plate number.

At this time, when the license plate number is recognized by the license plate recognition unit 37, the control unit 30 transmits the recognized license plate information to an external node.

Figure 11 is a flow chart showing the operation process of the rotation angle optimization unit of Figure 10.

The rotation angle optimization unit 38 of FIG. 11 is executed at every preset period (T) under the control of the control unit 30.

In addition, as shown in FIG. 11, the operation process (S38) of the rotation angle optimization unit 38 includes a polarization filter rotation angle initialization step (S381), a rotation control step (S382), a shooting control step (S383), and image analysis. And a license plate extraction step (S384), a light-saturation detection step (S385), a rotation final determination step (S386), a rotation angle increment calculation step (S387), and a light-saturation-based rotation angle optimal value determination step (S388). ).

The polarization filter rotation angle initialization step (S381) is a step of initializing the rotation angle (θ) of the polarization filter 75 to 0°.

At this time, the rotation range of the polarizing filter is from 0° to 90°, and the initial value of the rotation angle (θ) is preset to ‘0°’.

The polarization filter rotation control step (S382) is a step of controlling the filter rotation control unit 77 so that the polarization filter 75 is rotated at the rotation angle (θ = 0°) initialized by the polarization filter rotation angle initialization step (S381). .

The photographing control step (S383) is a step of controlling the camera unit 7 to photograph the license plate when the polarizing filter 75 is rotated by the polarizing filter rotation control step (S382).

In the image analysis and license plate extraction step (S384), when a photograph is taken by the camera unit 7 through the photographing control step (S383), the image acquired by the camera 7 is analyzed to obtain a license plate image. This is the extraction stage.

The light-saturation detection step (S385) is a step of detecting the light saturation of the license plate image extracted by the image analysis and license plate extraction step (S384).

The rotation final determination step (S386) is a step of determining whether the current rotation angle (θ) of the polarization filter 75 is '90°'. 1) If the current rotation angle (θ) of the polarization filter 75 is '90°', If it is '90°', the next step is to determine the optimal rotation angle value based on light-saturation (S388). 2) If the current rotation angle (θ) of the polarizing filter 75 is not '90°', the next step is The next step is to calculate the rotation angle increase value (S387).

The rotation angle increase value calculation step (S387) is performed when, in the rotation final determination step (S386), it is determined that the current rotation angle (θ) of the polarization filter is not '90°', and the current rotation angle of the polarization filter (75) is determined. The next rotation angle (θ', θ' = θ + α) of the polarization filter 75 is calculated by adding up the preset increase value (α) from the angle (θ).

In addition, in the rotation angle increase value calculation step (S387), when the next rotation angle (θ') of the polarization filter 75 is calculated, the next step returns to the above-mentioned polarization filter rotation control step (S382) and repeats the subsequent process.

At this time, when the polarization filter rotation control step (S382) is performed after the rotation angle increase value calculation step (S387), the polarization filter 75 is adjusted to the next rotation angle (θ') calculated in the rotation angle increase value calculation step (S387). The filter rotation control unit 77 is controlled to rotate.

The light-saturation-based rotation angle optimal value determination step (S388) analyzes the light-saturation values for each rotation angle of the polarization filter 75, calculates the rotation angle (θ) with the lowest light saturation among them, This is the step of determining the calculated rotation angle (θ) as the optimal rotation angle value.

At this time, when the optimal rotation angle value is determined through the light-saturation-based rotation angle optimal value determination step (S388), the control unit 30 outputs the determined rotation angle optimal value to the filter rotation control unit 77 of FIG. 2 described above, and filter The rotation control unit 77 rotates the polarization filter 75 according to the optimal rotation angle value received from the controller 3.

In this way, the rotation angle optimization unit 38 optimizes the rotation angle of the polarization filter 75 suitable for the current environmental conditions (illuminance, etc.) at each preset period (T), thereby significantly increasing the accuracy and reliability of license plate recognition. do.

In this way, the license plate recognizer 1, which is an embodiment of the present invention, is configured so that the polarizing filter installed in front of the camera lens filters the vertical vibration wave component of the input optical image, thereby effectively preventing oversaturation and spot phenomenon of the license plate image. By removing it, the accuracy and reliability of license plate recognition can be significantly increased.

In addition, the license plate recognizer (1) of the present invention is configured so that the polarizing lens installed in front of the LEDs removes the vertical vibration wave component of the optical signal emitted from the LEDs, but passes the horizontal vibration wave component, thereby increasing the license plate recognition rate. It can be improved further.

In addition, the license plate recognition device 1 of the present invention further includes a solar filter installed in front of the polarizing filter with a camera unit, thereby effectively blocking sunlight and further increasing the license plate recognition rate.

In addition, the license plate recognizer (1) of the present invention is configured so that the polarizing filter of the camera part can rotate, and at the same time, the controller is configured to calculate the optimal rotation angle of the polarizing filter every preset period (T), thereby maximizing the number recognition rate compared to the same environment. You can do it.

1:Vehicle number recognition device 3:Controller
5: Lighting unit 7: Camera unit
30: Control unit 31: Memory
32: data input/output unit 33: shooting control unit
34: Lighting control unit 35: Image analysis unit
36: Vehicle license plate extraction unit 37: Number recognition unit
38: Rotation angle optimization unit 51: IR LED
53: Polarized lens 71: Camera
73: Camera lens 75: Polarizing filter
77: Filter rotation control unit 79: Solar filter
S38: Rotation angle optimization method S381: Polarization filter rotation angle initialization step
S382: Polarizing filter rotation control step S383: Shooting control step
S384: Image analysis and license plate extraction step
S385: Light-saturation detection step S386: Rotation final determination step
S387: Rotation angle increase value calculation step
S388: Light-saturation-based rotation angle optimal value determination step

Claims (7)

In the license plate recognition device for recognizing the license plate number of a detected vehicle:
The vehicle number recognition device is
A camera that acquires an optical image by photographing a detection vehicle, a polarizing filter installed in front of the lens of the camera to filter the vertical/horizontal vibration wave components of the input optical image, and a filter that controls the rotation of the polarizing filter. A camera unit including a rotation control unit;
A controller that recognizes the license plate number of the detected vehicle by analyzing the optical image obtained by shooting the camera unit;
It is executed under the control of the controller, and is installed in front of the LEDs that emit optical signals and the LEDs, removes the vertical/horizontal vibration wave components of the emitted optical signals, and passes the horizontal vibration wave components of the optical signals. It includes a lighting unit including a polarizing lens,
The controller is
a shooting control unit that controls the operation of the camera unit;
a lighting control unit that controls the operation of the lighting unit;
an image analysis unit that analyzes the optical image passing through the polarizing filter of the camera unit;
a license plate extraction unit that extracts a license plate image from an optical image with reference to the data analyzed by the image analysis unit;
A license plate recognition unit that recognizes the license plate number by analyzing the license plate image extracted by the license plate extraction unit;
It includes a rotation angle optimization unit that is executed at each preset period (T),
The operation method (S38) of the rotation angle optimization unit performed by the controller is
Step 381 (S381) of initializing the rotation angle (θ) of the polarization filter to 0°;
Step 382 (S382) of controlling the filter rotation control unit to rotate the polarizing filter at the rotation angle (θ = 0°) initialized in step 381 (S381);
When the polarization filter is rotated in step 382 (S382), controlling the camera unit to capture a license plate (S383);
When photography is performed by the camera unit through step 383 (S383), step 384 (S384) of extracting a license plate image by analyzing the obtained optical image;
Step 385 (S385) of detecting the light saturation of the license plate image extracted by step 384 (S384);
Step 386 (S386) of determining whether the current rotation angle (θ) of the polarization filter is '90°';
In step 386 (S386), the process proceeds when it is determined that the current rotation angle (θ) of the polarization filter is not '90°', and a preset increase value ( Step 387 calculates the next rotation angle (θ', θ' = θ + α) of the polarization filter 75 by adding α), then returns to step 382 (S382) and repeats the subsequent process. (S387);
In step 386 (S386), when the current rotation angle (θ) of the polarization filter is determined to be '90°', light-saturation values for each rotation angle of the polarization filter are analyzed, and among these, light- After calculating the rotation angle (θ) with the lowest degree of saturation, step 387 (S387) of determining the calculated rotation angle (θ) as the optimal value of the rotation angle of the polarization filter in the next cycle,
The filter rotation control unit
A license plate recognition device, wherein the optimal rotation angle value is not fixed but changes by rotating the polarization filter according to the optimal rotation angle value determined by the rotation angle optimization unit. delete The method of claim 1, wherein the lighting unit
A license plate recognition device further comprising a horizontal diffusion lens (CLA, Collimating Lens Array) that diffuses the horizontal component of the optical signal emitted from the LEDs. The method of claim 1, wherein the camera unit
A license plate recognition device further comprising a solar filter installed in front of the polarizing filter to filter sunlight. The license plate recognition device according to claim 1, wherein the LEDs are IR LEDs (Infrared LEDs). delete The vehicle number recognition device according to claim 1, wherein the rotation angle of the polarization filter ranges from 0° to 90°.