JP2000268291A - License plate recognition device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To prevent omissions of detection and recognition of a vehicle even when the vehicle continually enters a vehicle detection area for a license plate recognition device. SOLUTION: This license plate recognition device which is disclosed performs vehicle detection when a license plate is detected from a video of a vehicle and, is equipped with a plate segment processing part 4 which segment a plate image from the video every detection and successively stores it in plural plate image memories 5 and a character recognition processing part 6 which reads the plate image in order of memory from the plate image memories 5, cuts a character out of the read plate image and performs character recognition. Thus, processing of the plate segment processing part 4 and that of the character recognition processing part 6 are performed in parallel.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a license plate recognizing apparatus for reading a license plate of a running vehicle and recognizing the character information.

[0002]

2. Description of the Related Art A license plate of a running vehicle is read as one means for identifying an individual vehicle while the vehicle is running, for the purpose of collecting tolls on an expressway, managing parking lots, etc. There is a demand for an apparatus that identifies a vehicle by recognizing identification information such as a registered place, a vehicle type code, and a serial number described therein. Conventionally, as a device for reading a vehicle license plate and recognizing the character information, a device of a single-shot processing type and a device of a frame-by-frame processing type are known. Here, the single shot method starts a series of processes from detection of a license plate in an image to recognition of character information in response to a vehicle detection signal from a vehicle detector. When a license plate is detected from the video image, the character information recognition process is started. Regardless of which of these methods is used, the method of recognizing a license plate is to sequentially execute a series of processes of video input → license plate cutout processing → character recognition processing → recognition result output. This is generally done by:

FIG. 13 is a block diagram illustrating the configuration of a license plate recognizing apparatus of a single-shot processing system, and FIG. 14 is a flowchart showing a license plate recognition process in a single-shot processing system. As shown in FIG. 13, the license plate recognizing device of the single-shot processing method is roughly composed of a camera 41, a vehicle detector 42, and a license plate recognition processing unit 43. , An analog-digital (A / D) conversion processing section 44, a plate cutout processing section 45, a character recognition processing section 46, and an output processing section 47.

[0004] A camera 41 is provided in a vehicle passage or the like, and captures an image of a running vehicle. The vehicle detector 42 detects that the object is a vehicle based on the size, shape, and the like of the object traveling in a vehicle passage or the like. The license plate recognition processing unit 43 recognizes information on the detected license plate of the vehicle. Further, a license plate recognition processing unit 43
When the A / D conversion processing unit 44 receives the vehicle detection signal from the vehicle detector 42, the A / D conversion processing unit 44 sends an image capture signal to the camera 41, and accordingly, the video signal output from the camera 41 Is converted to a digital signal.
The plate cutout processing unit 45 includes an A / D conversion processing unit 44.
From the digital image output from the license plate area (plate image). The character recognition processing unit 46 cuts out characters from the cut out plate image and performs character recognition. The output processing unit 47 outputs the result of the character recognition processing performed by the character recognition processing unit 46 to an external display device or information processing device (not shown).

Next, with reference to FIG. 14, the operation of the license plate recognizing apparatus of the single shot type will be described. When the vehicle detector 42 detects a vehicle (step X1), it notifies the A / D conversion processing unit 44 of this. Upon receiving this notification, the A / D conversion processing unit 44 extracts the video signal from the camera 41, converts the video signal into a digital signal, and outputs the digital signal to the plate cutout processing unit 45. As a result, the plate cutout processing unit 45 performs a plate image cutout process from the received digital image (step X2).
Output plate image. When receiving the plate image (step X3), the character recognition processing unit 46 performs a character recognition process on the plate image (step X3).
4) The output processing unit 47 performs a process of outputting the recognition result of the character information to the outside (step X5). by this,
The external display device displays the character recognition result on a display screen. Further, the external information processing apparatus can perform processing such as billing for a toll on an expressway or a parking fee in a parking lot based on the recognition result of the character information. In addition,
In the one-shot processing method, the vehicle detection signal during the plate cutout processing or the character recognition processing is ignored.
The extraction of the video signal from No. 1 is not performed.

FIG. 15 is a block diagram illustrating the configuration of a license plate recognizing device of a frame processing system, and FIG.
9 is a flowchart showing license plate recognition processing in each frame processing method. As shown in FIG. 15, the license plate recognizing device of the processing method for each frame uses:
It is roughly composed of a camera 51 and a license plate recognition processing section 52. The license plate recognition processing section 52 further includes an analog-digital (A / D) conversion processing section 53, a plate cutout processing section 54, and a character recognition. It comprises a processing unit 55 and an output processing unit 56.

[0007] The camera 51 is provided in a vehicle passage or the like, and captures an image of a running vehicle. The license plate recognizing unit 52 recognizes information on the license plate of the imaged vehicle. Further, in the license plate recognition processing section 52, the A / D conversion processing section 53 takes in the video signal from the camera 51 and converts it into a digital signal in accordance with the video interrupt signal output from the camera 51 for each frame. . The plate cutout processing unit 54 performs a process of cutting out a license plate area (plate image) from the digital image output from the A / D conversion processing unit 53. The character recognition processing unit 55 extracts characters from the extracted plate image and performs character recognition. Output processing unit 5
Reference numeral 6 outputs the result of the character recognition processing performed by the character recognition processing unit 55 to an external display device or information processing device (not shown).

Next, with reference to FIG. 16, the operation of the license plate recognizing apparatus of the frame processing system will be described. Since the camera 51 constantly captures an image of the vehicle and outputs a video interrupt signal to the A / D conversion processing unit 53 for each frame, the A / D conversion processing unit 53 sets the camera every time the video interrupt signal is received. The video signal is extracted from the video signal 51, converted into a digital signal, and output to the plate cutout processing unit 45. As a result, in the plate cutout processing unit 54,
The plate image is cut out from the received digital image (step Y1), and the plate image is output. When receiving the plate image (step Y2), the character recognition processing unit 55 performs a character recognition process on the plate image (step Y3).
A process of outputting the recognition result of the character information to the outside is performed (step Y4). In the frame-by-frame processing method, vehicle detection is performed by license plate cutout processing. In this method, a video interrupt signal during the character recognition processing is ignored, so that a new video signal is not taken out from the camera 51.

Japanese Patent Application Laid-Open No. 2-193266 discloses a system similar to the above-described system, in which a plurality of vehicle entry detection sensors and a plurality of images of a vehicle corresponding to the vehicle entry detection of each vehicle entry detection sensor are taken. License plate that outputs a determination result with respect to a vehicle having a plurality of lanes by providing one camera device and one image processing device having one image processing operation unit for a plurality of original image memories corresponding to each camera output A reader is disclosed.

[0010]

However, the conventional license plate recognition device has the following problems. The first problem is that recognition loss occurs when the vehicle continuously enters the vehicle detection range. This generally requires a certain amount of time for the character recognition process, but in the conventional license plate recognition device, the character recognition process is performed by sequential processing.
This is because, during the character recognition processing of the preceding vehicle, data of other vehicles that have continuously entered the detection range are ignored, and the processing is not performed. The second problem is that the license plate recognition rate decreases during traffic jams of vehicles. This is because the conventional license plate recognition device is a sequential processing method,
Processing is performed sequentially from the first captured image, but since character recognition processing takes time, the data captured during that processing is ignored, so the processing cannot be completed in time for vehicles entering one after another This is because the frequency of processing a good image including the entire license plate is extremely reduced.

[0011] The present invention has been made in view of the above circumstances, and in a license plate recognition device, even if a vehicle continuously enters a vehicle detection range,
It is an object of the present invention to prevent recognition loss from occurring and prevent the license plate recognition rate from being lowered even during traffic jams of vehicles.

[0012]

According to a first aspect of the present invention, there is provided a license plate recognizing apparatus for detecting a vehicle when a license plate is detected from an image of a vehicle. A plate cutout processing means for cutting out a plate image from the video for each time and sequentially storing the plate images in a plurality of plate image memories, reading out the plate images in the order of recording from the plate image memory, and cutting out characters from the read out plate image And a character recognition processing means for performing character recognition by performing a character recognition process, wherein the processing of the plate cutout processing means and the processing of the character recognition processing means are executed in parallel.

According to a second aspect of the present invention, there is provided the license plate recognizing device according to the first aspect, wherein the character recognizing processing means comprises a step of setting the character plate for each plate image corresponding to the earliest vehicle detection for each vehicle. It is characterized by performing recognition processing.

According to a third aspect of the present invention, there is provided a license plate recognizing device, in which a plate image is cut out from an image of the vehicle in response to a vehicle detection signal from a vehicle detector and is sequentially stored in a plurality of plate image memories. Plate cutting means, and character recognition processing means for reading plate images in the order recorded in the plate image memory, cutting out characters from the read plate images and performing character recognition, and processing the plate cutting processing means. And the processing of the character recognition processing means is performed in parallel.

According to a fourth aspect of the present invention, there is provided a license plate recognizing device, which detects a license plate from an image of a vehicle, detects the vehicle, and cuts out a plate image from the image for each vehicle detection. Plate cutout processing means for sequentially storing in a plurality of areas in the plate image holding memory for which character recognition has been completed among the plurality of plate image holding memories; and a plate for which character recognition has not been completed for all areas of the plurality of plate image holding memories. An image holding memory is selected, a plate image is held from a plurality of areas in the selected plate image holding memory, but an area for which the character recognition process has not been completed is selected, and the plate image is read. And character recognition processing means for performing character recognition by cutting out characters from the plate image. The processing of bets clipping processing means is characterized by being configured to process the process and the parallel of the character recognition processing means.

According to a fifth aspect of the present invention, there is provided the license plate recognizing device according to the fourth aspect, wherein the character recognizing processing section is the first of a plurality of areas of the selected plate image holding memory. It is characterized in that the character recognition processing is performed in order from the area where the plate image is stored.

The invention according to claim 6 relates to the license plate recognition device according to claim 4, wherein the character recognition processing unit confirms that the vehicle has passed the detection area,
It is characterized in that, from among the plurality of areas of the selected plate image holding memory, the character recognition processing is performed in order from the area where the plate image is stored last.

According to a seventh aspect of the present invention, there is provided the license plate recognizing device according to the fifth or sixth aspect, wherein the character recognizing processing unit performs processing on the plate image in any area of the selected plate image holding memory. After the character recognition process is performed, the plate image in the remaining area of the plate image holding memory is not subjected to the character recognition process.

[0019]

According to the structure of the present invention, the license plate recognizing device detects the license plate from the image of the vehicle, detects the vehicle, extracts a plate image from the image for each vehicle detection, and stores the plate image in a plurality of plate image memories. A plate cutout processing means for sequentially storing the plate image, and a character recognition processing means for reading out the plate images in the order of recording from the plate image memory, cutting out characters from the read plate image and performing character recognition, and Since the processing of the processing means and the processing of the character recognition processing means are performed in parallel, even if the vehicle continuously enters the vehicle detection range, the detection and recognition of the vehicle do not leak. Also, even during a traffic jam of the vehicle, the recognition rate of the license plate does not decrease.

[0020]

Embodiments of the present invention will be described below with reference to the drawings. The description will be specifically made using an embodiment. FIG. 1 is a block diagram showing a configuration of a license plate recognition apparatus according to a first embodiment of the present invention. FIG. 2 is a flowchart showing a license plate recognition process of the embodiment.
5 is a timing chart for explaining the operation of the embodiment. The license plate recognition device of this example is shown in FIG.
As shown in FIG. 1, the camera 1 and the license plate recognition processing unit 2 are schematically configured. The license plate recognition processing unit 2 further includes an analog-digital (A / D)
It comprises a conversion processing unit 3, a plate cutout processing unit 4, a plate image holding memory 5, a character recognition processing unit 6, and an output processing unit 7.

The camera 1 is provided in a vehicle passage or the like, and captures an image of a running vehicle. The license plate recognition processing unit 2 recognizes information on the license plate of the imaged vehicle. At this time, in the license plate recognition processing unit 2, the A / D conversion processing unit 3 reads a video signal from the camera 1 and converts it into a digital image according to a video interrupt signal output from the camera 1 for each frame. .
The plate cutout processing unit 4 cuts out the area of the license plate (plate image) from the digital image output from the A / D conversion processing unit 3 and performs vehicle detection by cutting out the plate image. The plate image holding memory 5 sequentially records the plate images cut out from the plate cutting processing unit 4. The plate image holding memory 5 can hold a plurality of images. The character recognition processing unit 6 monitors the image recording state of the plate image holding memory 5, sequentially reads the recorded plate images, cuts out characters from the read plate images, and performs character recognition. The output processing unit 7 outputs the result of the character recognition processing performed by the character recognition processing unit 6 to an external display device, information processing device, or the like (not shown).

Next, the operation of the license plate recognition device of this example will be described with reference to FIG. Camera 1
Always captures an image of a passing vehicle and outputs a video interrupt signal to the A / D conversion processing unit 3 for each frame, so that the A / D conversion processing unit 3 1 is converted to a digital signal and output to the plate cutout processing unit 4. This starts a plate image cutout loop, and the plate cutout processing unit 4 uses the received digital image to obtain a license plate area (plate image).
Is performed (step S1). When a plate image is cut out (step S2), the plate data is sequentially recorded in the plate image holding memory 5 (step S3). If the plate data has not been cut out, the cutout loop ends.

On the other hand, the character recognition processing section 6 performs the character recognition processing in parallel with the processing of cutting out and recording the plate data. At the start of the recognition loop, the character recognition processing unit 6
Checks whether there is plate data in the plate image holding memory 5 and, if there is plate data (step S11), performs character recognition processing (step S1).
2). When the character recognition processing is completed, the output processing unit 7
The recognition result in the character recognition processing unit 6 is output to an external display device or information processing device (not shown) (step S13). If there is no plate data in the plate image holding memory 5 (step S11),
End the recognition loop.

In this example, the plate image is cut out by, for example, arranging a character group indicating a registered place or a vehicle type code and a character group indicating a vehicle number, which are arranged in two rows in the plate image. This is performed by a method such as detecting the coincidence between the pattern and a pre-registered pattern to determine the area of the plate image, and at the same time, detecting the vehicle by the successful clipping of the plate image. If the same vehicle is detected a plurality of times within the vehicle detection range, a character recognition process of the plate image is performed based on the first vehicle detection. The first vehicle detection for the same vehicle means that, for example, by setting the angle of view of the camera 1, the vehicle detection range is limited to a range where only one vehicle is detected, and the name plate position in the image is changed. Using the fact that the nameplate position in the image changes to a higher position by utilizing the fact that the position decreases as the vehicle approaches, it can be recognized by determining the next vehicle. The character recognition processing is performed by, for example, cutting out a character area at a predetermined position in the area of the plate image, reading a character pattern, and detecting a match with a previously registered character pattern. The recognition processing is performed for all the characters of.

FIG. 3 explains the difference between the recognition process of the license plate recognition device of this example and the recognition process of the conventional license plate recognition device. In the figure, (1)
Shows vehicle detection in the case of the conventional single-shot processing method. Since vehicle detection is performed only once for each vehicle, for example, the first vehicle and the second vehicle 1, indicating that vehicle detection is performed. In the figure, (2) shows the conventional vehicle detection of each frame processing method, in which vehicle detection is performed for each vehicle for each frame. For example, the first vehicle and the first vehicle are detected. Vehicle 2, the first vehicle 3, the second vehicle 1, the second vehicle 2, the second vehicle 3, etc.
It is shown that vehicle detection is performed every frame.

(3) shows the recognition processing timing of the single-shot processing method. Since the sequential processing is performed, the recognition processing including the character recognition is performed by the vehicle detection of the first vehicle 1. Even if the vehicle detection of the second vehicle 1 is performed during the operation, the recognition process for the first vehicle is not completed, so that the recognition process for the second vehicle may be performed. No, it is shown that a missed detection occurs for the second vehicle. (4) in the figure shows the recognition processing timing of each frame processing method. Similarly, since the sequential processing is performed, the recognition processing is performed by detecting the first vehicle 1. Between two
Even if the vehicle detection is performed on the second vehicle, the recognition process on the first vehicle is not completed, so that the recognition process on the second vehicle cannot be performed.
It is shown that a detection leak occurs in the first vehicle.

(5) in the figure shows the vehicle detection by the license plate recognition device of this example, and the vehicle detection is performed for each vehicle every time plate data is cut out. For example, for the first vehicle, vehicle detection is performed as in the first vehicle 1, the first vehicle 2, the first vehicle 3, and so on. It is shown that vehicle detection is performed as in vehicle 1, second vehicle 2, second vehicle 3, and so on. (6) in the figure
This figure shows the recognition processing timing of the license plate recognition device of this example, and the recognition processing is performed for a plate image for which it is first detected that there is plate data for each vehicle. Thus, for example, the character recognition process is performed on the first vehicle based on the plate data recorded according to the vehicle detection of the first vehicle 1, and during the process, the first vehicle 2, 1 The vehicle detection of the third vehicle 3 is ignored. Then, after the recognition processing for the first vehicle is completed, the plate data recognition processing for the second vehicle based on the vehicle detection of the second vehicle 1 read from the plate image memory 5 is performed. This shows that no missed detection of the second vehicle occurs.

As described above, in the license plate recognizing device of this example, the vehicle detection is performed by the process of cutting out the plate image of the vehicle that has entered the detection range, and a plurality of plate image holding memories are provided. All images are recorded every frame, and plate cutout processing and character recognition processing are performed in parallel, so plate images of vehicles that have entered the detection range during the character recognition processing of the preceding vehicle are also recorded. After completion of the character recognition processing of the preceding vehicle, the character can be read out and character recognition can be performed. Therefore, even if the vehicle continuously enters, there is no omission in detection and recognition.

Second Embodiment FIG. 4 is a block diagram showing the configuration of a license plate recognition device according to a second embodiment of the present invention. The license plate recognition device of this example includes a camera 1 as shown in FIG.
1, a vehicle detector 12, and a license plate recognition processing unit 13. The license plate recognition processing unit 13 further includes an analog-digital (A /
D) Conversion processing unit 14 and plate cutout processing unit 15
, Plate image holding memory 16 and character recognition processing unit 1
7 and an output processing unit 18. The license plate recognition device of this example is significantly different from the license plate recognition device of the first embodiment shown in FIG. 1 in that a vehicle detector 12 is provided. The function of the / D conversion processing unit 14 is the same as that of the first embodiment.
This is different from the / D conversion processing unit 3.

The vehicle detector 12 detects that a vehicle has entered based on the size, shape, and the like of an object traveling in a vehicle passage or the like. The license plate recognition processing unit 13 recognizes information on the detected license plate of the vehicle. At this time, in the license plate recognition processing unit 13, when the A / D conversion processing unit 14 receives the vehicle detection signal from the vehicle detector 12, the A / D conversion processing unit 14 transmits an image capture signal to the camera 11 and responds accordingly. To convert the video signal output from the camera 11 into a digital signal. Other functions of the plate cutout processing unit 15, the plate image holding memory 16, the character recognition processing unit 17, and the output processing unit 18 are the same as those of the plate cutout processing unit 4, plate image processing in the first embodiment shown in FIG. This is the same as the holding memory 5, the character recognition processing unit 6, and the output processing unit 7.

In the license plate recognition device of this example,
As in the case of the first embodiment, the processing of the plate data cutout loop and the processing of the plate data recognition loop are executed by parallel processing. In the cutout loop, when the vehicle detector 12 detects a vehicle, the vehicle detector 12 notifies the A / D conversion processing unit 14 of the detection, and the A / D conversion processing unit 14 extracts a video signal from the camera 11 and outputs a digital signal. And outputs it to the plate cutout processing unit 15. The plate cutout processing unit 15 performs a plate image cutout process from the received digital image. When the plate image is cut out, the plate data is sequentially converted to the plate data. It is recorded in the image holding memory 16. On the other hand, in the recognition loop, when the character recognition processing unit 17 performs the character recognition processing in parallel with the plate cutout processing and the plate data recording processing, and detects that there is plate data in the plate image holding memory 16, the character recognition processing unit 17 performs the character recognition. The processing is performed, and the output processing unit 18
The result of character recognition in step 7 is output to an external display device or information processing device (not shown).

As described above, in the license plate recognition device of this example, vehicle detection is performed by vehicle detection of the vehicle detector, and a plurality of plate image holding memories are provided so that the vehicle can be detected within the detection range during character recognition processing. The plate images of the vehicles that entered were recorded one by one for each vehicle, and the plate cutout process and the character recognition process were performed in parallel. The plate image of the vehicle that has entered the vehicle is also recorded, and the character can be read out and character recognition can be performed after the character recognition processing of the preceding vehicle has been completed. Therefore, even when the vehicle continuously enters, detection and recognition can be performed. Leakage can be prevented, and the capacity of the plate image holding memory can be reduced as compared with the first embodiment.

Third Embodiment FIG. 5 is a block diagram showing the configuration of a license plate recognition device according to a third embodiment of the present invention. FIG. 6 shows the relationship between the vehicle position and the memory area in the third embodiment. FIG. 7 is a flowchart showing the operation of the memory management processing unit in the embodiment, FIG. 8 is a flowchart (1) showing the operation of the character recognition processing unit in the embodiment, and FIG. Flowchart (2) showing the operation of the character recognition processing unit, FIG. 10 is a flowchart (3) showing the operation of the character recognition processing unit in the embodiment, and FIG. 11 shows that the character recognition process is performed in the preceding memory area. FIG. 12 is a timing chart showing an operation when a character recognition process is performed in the next memory area.

As shown in FIG. 5, the license plate recognition apparatus of this example is roughly constituted by a camera 21 and a license plate recognition processing section 22. Further, the license plate recognition processing unit 22 includes an analog-digital (A / D) conversion processing unit 23, a plate cutout processing unit 24, a memory management processing unit 25, and a first plate image holding memory (memory A) 26. A second plate image holding memory (memory B) 27, a character recognition processing unit 28,
An output processing unit 29 is provided. The license plate recognition device of this example is different from the first embodiment shown in FIG.
This is greatly different from the point that it has two plate image holding memories 26 and 27, and accordingly, the function of the character recognition processing unit 28 is different from that of the character recognition processing unit 6 of the first embodiment.

The memory management processing unit 25 manages recording of a plurality of plate images from the camera 21 in the two plate image holding memories 26 and 27. The plate image holding units 26 and 27 are provided corresponding to the number of detectable vehicles 2, and respectively hold a plurality (for example, three) of plate images. The character recognition processing unit 28 includes a memory management processing unit 25
Then, according to the plate image recording states of the plate image holding memories 26 and 27 and the current attention memory information, any one of the plate images in the plate image holding memories 26 and 27 is read and character recognition processing is performed. Other than this, A
The functions of the / D conversion processing unit 23, the plate cutout processing unit 24, and the output processing unit 29 are the same as those of the A / D conversion processing unit 3, the plate cutout processing unit 4, and the output in the case of the first embodiment shown in FIG. This is the same as the processing unit 7.

Next, the operation of the memory management processing unit of this example will be described with reference to FIG. 6, reference numeral 21 denotes a camera, and reference numerals 26 and 27 denote plate image holding memories (A) and (B), respectively, which correspond to the configuration shown in FIG. 31 is a detection (imaging) range, 3
2 is an input image from the camera 21 in the detection range, 33
Is the vehicle passing the nth vehicle, and 34 is the next (n + 1) of the vehicle 33
This is a vehicle that passes through the vehicle. W1, w of input image 32
2 and w3 indicate areas of areas 1, 2, and 3, respectively, and correspond to areas W1, W2, and W3 of the detection range 31, respectively.

In this example, while one vehicle passes through the detection range, the number of vehicles holding plate data and the number of images are limited, and images that do not need to be recognized are discarded. When the vehicle n enters the first area W1 of the detection range 31, the memory management processing unit 25
When a plate is detected in the range of 1, the recording state of the plate image holding memory A is cleared, and the plate image at that time is recorded in the area A1 of the plate image holding memory A,
The recording state of the plate image holding memory A is updated. If a plate is detected in the range of w1 in the next input image, the plate image at that time is not recorded. Similarly, when the vehicle n enters the area W2 next to the detection range 31 and the plate of the vehicle n is detected for the first time in the range w2 of the input image, the plate image at that time is stored in the plate image holding memory. A is recorded in the area A2 of A, and the recording state of the plate image holding memory A is updated. If a plate is detected in the range of w2 in the next input image, the plate image at that time is not recorded. Similarly, when the plate of the vehicle n is detected for the first time in the range of w3 of the input image due to the vehicle n entering the area W3 of the detection range 31, the plate image at that time is stored in the plate image holding memory A. Recorded in area A3, plate image holding memory A
Update the recording status of. Plate is w3 in the next input image
, The plate image at that time is not recorded.

When the next vehicle n + 1 enters the detection range 31 and detects a plate in the input image 32, the same operation as described above is performed on the plate image holding memory B.

The memory management processing section 25 enables the character recognition processing section 28 to determine whether the images recorded in the plate image holding memories A and B are new ones or those of vehicles that have passed in the past. Therefore, information indicating the plate image holding memory of interest (to be written in response to the next input), that is, information on whether the plate image holding memories A or B are transmitted to the character recognition processing unit 28. .

Hereinafter, the flow of processing in the memory management processing unit of this example will be described with reference to FIG. The memory management processing unit 25 detects that the plate has been cut out by the plate cutout processing unit 24 (step P
1) If it is recognized that it belongs to a new vehicle (step P2), the plate image holding memories A and B
Is performed (step P3), and the recording state of the switched memory is cleared (step P4). When the vehicle is in area 1 of the input image corresponding to area W1 (step P5), when the vehicle is not already recorded in area 1 of the plate image holding memory (step P6), the vehicle is recorded in area 1 of the plate image holding memory. Then, the recording state of the memory is updated (step P7). If the data has already been recorded in area 1 (step P6), the recording process is not performed. If the vehicle is not in area 1 of the input image (step P5), but is in area 2 (step P8), but not already recorded in area 2 of the plate image holding memory (step P9), area 2 of the plate image holding memory
To update the recording state of the memory (step P
10). If the data has already been recorded in the area 2 (step P9), the recording process is not performed. Furthermore, when the vehicle is not in area 1 of the input image (step P
5) If not in area 2 (step P8), if not already recorded in area 3 of the plate image holding memory (step P11), record in area 3 of the plate image holding memory and update the recording state of the memory (Step P12). If the data has already been recorded in the area 3 (step P11), the recording process is not performed.

Next, the operation of the character recognition processing unit in this example will be described. The character recognition processing unit 28 receives, from the memory management processing unit 25, the recording states of the plate image holding memories 26 and 27 and the information of the current memory of interest. When there is no plate data, the character recognition processing unit 28 does not perform the operation. If there is plate data and all the characters of the plate of the vehicle recorded in the plate image holding memory to which the memory management processing unit 25 is not paying attention are not recognized, the memory management processing unit 25 pays attention. The plate image holding memory which does not exist is set as a recognition target memory. At this time, if all the characters have already been recognized, the plate image holding memory to which the memory management processing unit 25 focuses is set as a recognition target. If all areas of the plate image holding memory to be recognized have already been subjected to character recognition processing, the plate image holding memory to be recognized is cleared and the plate image holding memory to be recognized is changed. Plate images are sequentially recorded in areas 1 to 3 of the plate image holding memory to be recognized, and an area in which the character recognition process is not performed is searched for. If detected, the area is recorded in that area. The character recognition processing of the plate image is performed.

The flow of processing in the character recognition processing unit of this example will be described below with reference to FIGS.
As shown in the flowchart of FIG. 8A, when the recognition loop of the character recognition processing unit 28 starts, it is determined whether there is plate data cut out from the plate cutout processing unit 24. Step Q1) In the memory management processing unit 25, it is checked whether or not the plate image holding memory A is focused (step Q2). If the plate image holding memory A is not focused, the flow proceeds to the flow of FIG. When attention is paid to the image holding memory A, the flow proceeds to the flow of FIG.

As shown in the flowchart of FIG. 9, in the plate image holding memory A, if all the characters of the plate image have not been recognized (step R1), and if not all the areas have been recognized (step R2), It is determined whether or not there is data in the memory A1 of the plate image holding memory A (step R3). If there is plate data in the memory A1 and the recognition processing of the memory A1 has not been completed (step R4), the character recognition processing is performed. (Step R8),
Moving to the processing of FIG. 8B, output processing is performed (step V1), and the recognition loop ends. When there is no plate data in the memory A1 (step R3), and when the recognition processing of the memory A1 has been completed (step R4), it is determined whether or not there is plate data in the memory A2 (step R3).
5) If there is plate data in the memory A2 and the memory A2 has not been processed (step R6), a character recognition process is performed (step R8), and the process proceeds to the process of FIG. (Step V1) to end the recognition loop. When there is no plate data in the memory A2 (step R5) and when the recognition processing of the memory A2 has been completed (step R6), it is determined whether there is plate data in the memory A3. (Step R7), the character recognition process is performed (Step R8), and the process proceeds to the process of FIG. 8B, the output process is performed (Step V1), and the recognition loop is ended. If there is no plate data in the memory A3 (step R
7), the process proceeds to the flow of FIG.
When all the characters of the plate image have been recognized in the plate image holding memory A (step R1), the memory A is cleared (step R9), and the process proceeds to the flow of FIG. If not all characters of the plate image in the memory A have been recognized (step R1), but all areas have been recognized (step R2), the process proceeds to the flow of FIG.

As shown in the flowchart of FIG. 10, if all the characters of the plate image have not been recognized in the plate image holding memory B (step U1) and if the recognition processing of all the areas has not been completed (step U2), It is determined whether or not there is data in the memory B1 of the plate image holding memory B (step U3). If there is plate data in the memory B1 and the recognition processing of the memory B1 has not been completed (step U4), the character recognition processing is performed. (Step U
8), the process proceeds to the process of FIG. 8B, an output process is performed (step V1), and the recognition loop ends. When there is no plate data in the memory B1 (step U3) and when the recognition processing of the memory B1 has been completed (step U4), it is determined whether or not there is plate data in the memory B2 (step U5). There and
If the memory B2 has not been processed (step U6),
The character recognition process is performed (step U8), and the process proceeds to the process of FIG. 8B, the output process is performed (step V1), and the recognition loop ends. When there is no plate data in the memory B2 (step U5), and when the recognition processing of the memory B2 has been completed (step U6), it is determined whether there is plate data in the memory B3. (Step U7), the character recognition process is performed (Step U8), and the process proceeds to the process of FIG. 8B, the output process is performed (Step V1), and the recognition loop ends. If there is no plate data in the memory B3 (step U
7), the process proceeds to the flow of FIG.
When all the characters of the plate image have been recognized in the plate image holding memory B (step U1), the memory B is cleared (step U9), and the process proceeds to the flow of FIG. Even if all the characters of the plate image in the memory B have not been recognized (step U1), but all the areas have been recognized (step U2), the process proceeds to the flow of FIG.

By such an operation, the memory in the preceding memory
The process is performed sequentially from the area, and for a vehicle that once recognizes all the characters, the character recognition process is not performed for a plate image cut out thereafter. This makes it possible to prevent a plate image from being missed based on the character recognition processing time.

FIG. 11 explains the difference between the recognition processing of the license plate recognition apparatus of this example and the recognition processing of the conventional license plate recognition apparatus when character recognition processing is performed in the preceding memory area. Is what you do.

In the figure, (1) shows the license plate recognition apparatus of this example and the conventional vehicle detection of each frame processing method, and the vehicle detection is performed for each vehicle every time a plate image is cut out. For example, for the first vehicle, 1-1, 1-2, 1-3, 1-4, 1
Vehicle detection is performed as in -5, 1-5, 1-6, 1-7, 1-8, and 1-9.
It is shown that the vehicle detection is performed as in 2-2, 2-3, 2-4, 2-5, 2-6. (2) shows the recognition processing timing of each frame processing method, and the recognition processing is started for the first vehicle based on 1-1 vehicle detection. 9 immediately before the vehicle detection of the ninth vehicle, the second recognition process is started for the first vehicle based on the vehicle detection of 1-9, and the process is performed by the second vehicle 2-1 and 2-2-1. 2,2-
This indicates that the recognition process was not performed on the second vehicle because the vehicle detection did not end within the period of vehicle detection of 3, 2-4, 2-5, and 2-6. (3) shows the recognition processing timing of the license plate recognition device of this example. The recognition processing was performed on the first vehicle based on the 1-1 vehicle detection. As for the vehicle detection of (1), since it is recognized that the same vehicle is detected, the processing is not performed. Further, it is shown that the recognition processing of the second vehicle is performed based on the vehicle detection of 2-1.

In the license plate recognition device of this example, if all characters cannot be recognized in the preceding area, the character recognition processing can be performed in the next area. It is possible to prevent the recognition rate from lowering.

FIG. 12 explains the difference between the recognition processing of the license plate recognition apparatus of this example and the recognition processing of the conventional license plate recognition apparatus when the character recognition processing is performed in the next memory area. Is what you do. In the figure,
(1) shows the vehicle detection in the case of the license plate recognition apparatus of this example and the conventional frame-by-frame processing method. Vehicle detection is performed for each vehicle every time a plate image is cut out. For the second vehicle, vehicle detection is performed as in 1-1, 1-2, 1-3, 1-4, and 1-5, and for the second vehicle, 2-1 and 1-2.
It is shown that vehicle detection was performed like -2, 2-3, 2-4, 2-5. (2) shows the recognition processing timing of each frame processing method. In the first vehicle, an error occurred in the recognition processing performed based on the 1-1 vehicle detection. Since the recognition process is not performed for the first vehicle after the recognition process time, the process directly proceeds to the recognition process for the second vehicle.
Therefore, it is shown that the recognition processing was not performed for the first vehicle. (3) shows the recognition processing timing of the license plate recognition device of this example, and is based on the detection of 1-1 vehicle of the first vehicle.
Although an error occurred in the first recognition processing, the second recognition processing based on the 1-2 vehicle detection of the first vehicle is performed, and the recognition processing of the first vehicle succeeds. 2 shows that the recognition processing based on the vehicle detection of 2-1 has been performed on the second vehicle.

As described above, in the license plate recognizing device of this example, vehicle detection is performed by cutting out a plate image of a vehicle that has entered the detection range, and a plate image composed of a plurality of areas is provided for each of a plurality of vehicles. An image holding memory is provided to record the cut-out plate image and perform the plate cut-out process and the character recognition process in parallel, so that the vehicle that has entered the detection range during the character recognition process of the preceding vehicle is provided. The plate image is also recorded, and after the character recognition processing of the previous vehicle is completed, it can be read out and character recognition can be performed. Therefore, even if the vehicle enters continuously, leakage of detection and recognition may occur. There is no vehicle, and once all characters have been recognized,
Since the character recognition processing of the clipped plate image is not performed thereafter, it is possible to prevent the plate image from being missed based on the character recognition processing time. Further, when all characters cannot be recognized in the preceding area, character recognition can be performed in the next area, so that a reduction in recognition rate can be prevented.

Although the embodiment of the present invention has been described in detail with reference to the drawings, the specific configuration is not limited to this embodiment, and there are design changes and the like without departing from the gist of the present invention. Even this is included in the present invention. For example, in the third embodiment, the configuration of the plate image holding memory is the number of vehicles 2 and the number of images 3. However, the configuration is not limited to this, and the number of vehicles and the number of images may be larger. Further, in the third embodiment, each area of the plate image memory is processed in order from the front area. However, after confirming that the vehicle has passed the detection area, the processing is performed in order from the rear area. According to this method, since the license plate becomes closer to the camera in the rear area, the resolution is increased, and the accuracy of character recognition can be improved.

[0052]

As described above, according to the license plate recognizing device of the present invention, the plate cutting process and the character recognizing process are executed in parallel, so that the vehicles are continuously located within the vehicle detection range. Even when the vehicle enters, no leakage of the detection and recognition of the vehicle occurs. Further, when the character recognition is successful by any of the plate images in the plurality of areas, the character recognition processing is not performed on the remaining plate images, so that the plate image can be prevented from being lost due to the character recognition processing time, When the character recognition fails, the character recognition process is performed using the plate information of the next area, so that the license plate recognition rate does not decrease even during traffic jam of the vehicle.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a license plate recognition device according to a first embodiment of the present invention.

FIG. 2 is a flowchart showing a license plate recognition process of the embodiment.

FIG. 3 is a timing chart for explaining the operation of the embodiment.

FIG. 4 is a block diagram showing a configuration of a license plate recognition device according to a second embodiment of the present invention.

FIG. 5 is a block diagram showing a configuration of a license plate recognition device according to a third embodiment of the present invention.

FIG. 6 is a diagram showing a relationship between a vehicle position and a memory area in the embodiment.

FIG. 7 is a flowchart showing an operation of a memory management processing unit in the embodiment.

FIG. 8 is a flowchart (1) showing an operation processing procedure of a character recognition processing unit in the embodiment.

FIG. 9 is a flowchart (2) illustrating an operation processing procedure of the character recognition processing unit in the embodiment.

FIG. 10 is a flowchart (3) showing an operation processing procedure of the character recognition processing unit in the embodiment.

FIG. 11 is a timing chart showing an operation when a character recognition process is performed in a preceding memory area.

FIG. 12 is a timing chart showing an operation when character recognition processing is performed in a next memory area.

FIG. 13 is a block diagram illustrating an example of an electrical configuration of a license plate recognizing device of a single processing type.

FIG. 14 is a flowchart showing a license plate recognition processing procedure in a single shot processing method.

FIG. 15 is a block diagram illustrating a configuration of a license plate recognition device of a frame processing system.

FIG. 16 is a flowchart showing license plate recognition processing in each frame processing method.

[Explanation of symbols]

 1,11,21 camera 12 Vehicle detector 2,13,22 License plate recognition processing unit 3,14,23 A / D conversion processing unit 4,15,24 Plate cutout processing unit 5,16,26,27 Plate image holding Memory 6, 17, 28 Character recognition processing unit 7, 18, 29 Output processing unit 25 Memory management processing unit

Claims (7)

[Claims] 1. A plate cutout processing means for detecting a license plate from an image of a vehicle, performing a vehicle detection, cutting out a plate image from the image for each vehicle detection, and sequentially storing the plate image in a plurality of plate image memories; Character recognition processing means for reading out the plate images in the order of recording from the plate image memory, extracting characters from the read plate images and performing character recognition,
A license plate recognition apparatus characterized in that the processing of the plate cutout processing means and the processing of the character recognition processing means are executed in parallel. 2. The license plate recognition apparatus according to claim 1, wherein said character recognition processing means performs said character recognition processing for each plate image corresponding to the earliest vehicle detection for each vehicle. 3. Plate extracting means for extracting a plate image from an image of the vehicle in accordance with a vehicle detection signal from a vehicle detector and sequentially storing the plate image in a plurality of plate image memories; Character recognition processing means for reading out an image, cutting out characters from the read-out plate image and performing character recognition, and performs processing of the plate cutout processing means and processing of the character recognition processing means in parallel A license plate recognizing device characterized by being configured as described above. 4. A plate image which is detected when a license plate is detected from an image of a vehicle, and a plate image is cut out from the image every time the vehicle is detected, and a plate image of a plurality of plate image holding memories for which character recognition is completed. Plate cutout processing means for sequentially holding in a plurality of areas in the holding memory; and selecting a plate image holding memory for which character recognition of all areas has not been completed among the plurality of plate image holding memories, and selecting the selected plate image holding memory. A character recognition process in which a plate image is held from a plurality of areas inside but a character recognition process is not completed, a plate image is read out, characters are cut out from the read plate image, and character recognition is performed. Means, processing of the plate cutout processing means, processing of the character recognition processing means, Number plate recognition apparatus characterized by being configured to process in parallel. 5. The apparatus according to claim 1, wherein the character recognition processing unit is configured to perform the character recognition processing sequentially from an area in which the plate image is stored first among a plurality of areas of the selected plate image holding memory. The license plate recognition device according to claim 4, characterized in that: 6. The character recognition processing unit confirms that the vehicle has passed through the detection area, and sequentially starts with an area in which the plate image is stored last among a plurality of areas of the selected plate image holding memory. The license plate recognition device according to claim 4, wherein the license plate recognition device is configured to perform a character recognition process. 7. After the character recognition processing unit performs a character recognition process on a plate image in any area of the selected plate image holding memory, the character recognition processing unit performs processing on the plate image in the remaining area of the plate image holding memory. 7. The license plate recognition device according to claim 5, wherein the device is configured not to perform a character recognition process.