JPH07105352A - Picture processor - Google Patents

Abstract

PURPOSE:To provide the picture processor in which plural video cameras are used to process an object range in excess of a conventional picture range and a picture is processed by dividing the processing object area with the constitution simpler than heretofore at low cost. CONSTITUTION:This picture processing unit is provided with plural video cameras 11, 12 whose pickup ranges are partly overlapped respectively and a picture processing section 20 to which the video cameras are connected, and the plural video cameras 11, 12 store a picture by one pattern with an external shutter signal and provided an output of the stored picture with an external signal, a picture processing section 20 applies a series of operations such as sending of the shutter signal to each video camera in almost the same timing, receiving the stored picture from the video camera and processing the picture to each video camera sequentially till an objective picture processing result is obtained.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus, and more particularly to an image processing apparatus using a plurality of video cameras for processing a target range exceeding a normal image range with simultaneity.

[0002]

2. Description of the Related Art Image processing apparatuses for image-processing a video signal from a video camera are used in various fields. For example, even at the gate of the parking lot (see Figure 2) for automation,
It recognizes the vehicle registration numbers of vehicles entering and leaving by image processing.

In this type of image processing apparatus, a video signal including a vehicle, which is generally obtained by an image input section, is stored as a still image in an image storage section, and then a license plate extraction section extracts a license plate equivalent section. After the binarization processing section properly binarizes, the character cut-out section extracts the character equivalent sections, and the character recognition section character-recognizes the characters corresponding to these character equivalent sections. Note that each unit is generally configured as an integrated processing device.

Various types of devices for extracting such license plates have been proposed, for example, Japanese Patent Application No. 4-1377873 (image processing device) proposed by the present applicant. A license plate extraction method has also been proposed, for example, by the present applicant as Japanese Patent Application Laid-Open No. 5-143888 (Japanese Patent Application No. 3-332552).
Regarding the method of recognizing the vehicle registration number, for example, Japanese Patent Application No. 4-137874 (a method of recognizing numbers etc.) by the applicant
I have a suggestion.

When a vehicle registration number is recognized by using image processing as described above, a normal video camera is
If you use only a stand, the resolution will be insufficient if you set it with a field of view that covers the entire existing range of license plates, and the recognition of characters that display the Land Transport Bureau (kanji) and characters that display whether it is for business or not (hiragana). Is impossible.

As a countermeasure, it is conceivable to adopt a high resolution camera and a high resolution image processing unit corresponding thereto, but it is currently costly to use the high resolution camera and the high resolution image processing device easily. It is impossible for me.

In order to input a high resolution image to deal with this problem, there has already been proposed an apparatus for dividing a target range by using a plurality of video cameras to pick up and process the image. In addition, in the above-mentioned use, the simultaneity of processing is required. As such a device, for example, in Japanese Laid-Open Patent Publication No. 3-224100, there is a "license plate recognition device" that secures the visual fields of the left and right cameras by overlapping the central portions of the lane width in order to deal with deviation of the vehicle. It is disclosed. However, this proposal requires an image input section and an image memory for each of the left and right cameras.

In addition, Japanese Laid-Open Patent Publication No. 2-196400 discloses "vehicle recognition" in which a plurality of cameras are used to secure a field of view covering the entire road and character recognition of a license plate can be performed regardless of the traveling position of the vehicle. A device is disclosed. This proposal also requires the same number of image input units and image memories corresponding to each of the plurality of cameras.

[0009]

In the above-mentioned proposals and the like, since each camera is provided with the image input section and the image memory, the image processing apparatus side has a heavy load. That is, in order to perform image processing, it is usually necessary to save an input image and a processing process, and even if one video camera is supported, an image memory for saving a plurality of images must be provided. In addition to this, each camera must be equipped with an image input unit and an image memory, which makes the device complicated and expensive.

The respective inventions of the present application have been made by paying attention to the problems of the prior art as described above, and divide a processing target area by using a plurality of video cameras to perform image processing while maintaining simultaneity. It is an object of the present invention to propose an image processing device that significantly reduces the load on the image processing device side. The second invention of the present application proposes an image processing apparatus particularly suitable for vehicle registration number recognition.

[0011]

In order to achieve the above object, in the first invention of the present application, an image processing apparatus has a plurality of video cameras and an image processing section to which these video cameras are connected, The video camera stores an image for one screen according to a shutter signal from the outside, and outputs the stored image according to a selection signal from the outside. The image processing unit shutters the video cameras at substantially the same timing. Each unit is configured to sequentially output a signal, and then sequentially input a stored image from an individual video camera and perform image processing on each video camera until a predetermined image processing result is obtained. Constitute.

In the second invention of the present application, an image processing device is
A plurality of video cameras that share the vehicle traveling path in the width direction and share the images to obtain a foreground or a rear view of the traveling vehicle, and an image processing unit that recognizes the vehicle registration number from the outputs of these video cameras. And having the plurality of video cameras,
An image for one screen is stored by a shutter signal when the vehicle enters from the image processing unit, and the stored image is output to the image processing unit by a signal from the image processing unit. The shutter signal is sent to the video cameras at almost the same timing, then the stored images are output from each video camera, and the images selectively input through the camera switching unit are subjected to image processing up to the license plate extraction. The operation is sequentially performed on each video camera until the license plate portion is appropriately extracted, and after the license plate portion is extracted, each portion is configured to recognize the vehicle registration number.

[0013]

In both of the inventions, each video camera simultaneously stores an image for one screen in response to the input of a shutter signal from the outside (for example, an image processing section). Therefore, the screens of the individual video cameras have simultaneity and are zoomed up to form a high-resolution image for the processing target.

Next, the image processing section first inputs a stored image from one of the video cameras and performs predetermined image processing based on the input image. When the desired image processing result is not obtained, the stored image is input from the next video camera and the predetermined image processing is performed based on the input image. If the desired image processing result is still not obtained, the same operation is repeated. During this process, the above operation is stopped when the desired image processing result is obtained, and the image processing for one time is completed. By doing so, the image input unit and the image memory are shared, and the number of components as a whole is reduced.

Further, particularly in the image processing apparatus (vehicle registration number recognition apparatus) of the second invention of the present application for recognizing the vehicle registration number, each video camera corresponds to the input of the shutter signal based on the entry of the vehicle from the image processing unit. Then, the images for one screen, which partially overlap in the width direction of the vehicle traveling path, are simultaneously stored. Therefore, as a whole, the foreground or the foreground of the traveling vehicle can always be obtained, and the screen of each video camera is zoomed up to form a high-resolution image for the license plate portion to be processed. The image processing unit then sends a signal to cause the video camera to output the stored image, and first performs image processing up to at least the license plate portion extraction based on the image input from one of the video cameras.

If the license plate portion cannot be extracted, a signal is transmitted, the next stored image of the video camera is input, and the license plate portion extraction processing is similarly performed based on this. If the desired image processing result is not obtained, the same series of operations is repeated. During this process, the above operation is stopped when the license plate can be extracted. Then, subsequently, the vehicle registration number recognition processing is performed, and the image processing for one time is completed. After all, the configuration is simplified by sharing the image input unit and the image memory.

[0017]

Embodiments of the present invention will be described below in detail with reference to the drawings based on embodiments. FIG. 1 is a block diagram showing the configuration of an image processing apparatus (vehicle registration number recognition apparatus) which is an embodiment of the second invention of the present application. Further, FIG. 2 is a schematic plan view of the parking lot gate portion for explaining the arrangement and the like of the embodiment apparatus. The apparatus of this embodiment is also an embodiment of the image processing apparatus of the first invention of the present application.

The image processing device of the embodiment is a device for recognizing a vehicle registration number of a vehicle entering and leaving a parking lot, and is installed beside the vehicle running path 6 at the parking lot gate. This vehicle registration number recognition device includes two (plural) video cameras 1.
1 and 12, and an image processing unit (recognition unit) connected to this
Equipped with 20. In addition, a vehicle head sensor including a light projector 31A and a light receiver 31B and a strobe illuminator 32 are provided and connected to the image processing unit 20, respectively.

In this device, the strobe emits light when the vehicle head sensor detects an approaching vehicle. The two video cameras are
The shutter operation is performed within the strobe light emission time, and the images for one screen are stored simultaneously with each other. CPU of image processing unit
And the image processor takes in the stored image of the first video camera and performs image processing to determine whether the license plate is completely present. If it is present, the license plate is extracted, and then character recognition processing is performed. If not, the image stored in the remaining video camera is taken in and the same processing as above is performed. Hereinafter, the configuration of each unit and each processing process will be described in detail.

The two video cameras 11 and 12 are designed so as to partially overlap in the width direction of the vehicle traveling path and share images to obtain the foreground of the traveling vehicle. That is, the video camera 11 is installed so as to share the image pickup area R1 and the video camera 12 is arranged so as to share the image pickup area R2 to obtain a corresponding image (see FIG. 4).

Both video cameras are configured to store an image for one screen by a shutter signal from the outside (image processing unit 20 in the embodiment). Further, the stored image is output to the image processing unit as a video signal by a selection signal from the outside (image processing unit).

The image processing unit 20 includes a camera switching unit 21 to which each video signal corresponding to each stored image described above is input, an image input unit 22 sequentially connected to the camera switching unit 21,
An image memory 23, an image processor 24 connected to the image memory 23 by a direct bus, and a CPU 25 connected to the above respective units via a CPU bus to control the respective units and input / output data, the CPU 2
5 has an I / O circuit 26 which is also connected to the CPU bus for inputting / outputting with the outside.

The I / O circuit 26 receives a vehicle head detection signal from the light receiver 31B which constitutes a vehicle head sensor together with the light projector 31A. The I / O circuit 26 outputs the above-mentioned shutter signal to the video camera 11 and the video camera 12, and also outputs a strobe control signal for driving the strobe illuminator 32. Further, the vehicle registration number recognition result, which is the final result of the image processing, etc., is also transmitted to a remote host device via this I / O circuit 26.

The camera switching section 21 is provided for both video cameras 1.
The image input unit 2 to which the video signals from 1 and 12 are input and which is followed by the video signal selected and switched under the control of the CPU 25
Send alternatively to 2. The image input unit 22, the image memory 23, the image processing processor 24 and the like are well-known parts, and are subjected to predetermined image processing under the control of the CPU 25, that is, license plate portion extraction and character extraction from this extraction portion,
Processing such as recognition of each character is performed. As described above, as a result of sharing the image input unit and the image memory corresponding to a plurality of video cameras, the apparatus of the embodiment has a simple structure without sacrificing the processing time, and thus becomes a low-cost apparatus. .

Next, the operation of the embodiment apparatus will be described. A light beam is constantly emitted from the light projector 31A toward the light receiver 31B. The light receiver 31B transmits a signal different from that at the time of light reception (normal state) in response to the light from the light projector 31A being blocked by the entering vehicle. This change (rise or fall) is recognized by the image processing unit 20 as a vehicle head detection signal.

When the vehicle head detection signal is input, the image processing section 20 outputs a strobe control signal to the illumination controller 33 to drive the strobe illuminator 32, and at the same time, the video camera 11 and the video camera 12 capture images. Camera control signal S1 (for the shutter signal,
Hereinafter, also referred to as an image capture signal) is transmitted at substantially the same timing.

FIG. 3 is a diagram for explaining an example of the timing of each of the above signals. The strobe control signal shown in (b) is transmitted in response to the vehicle head detection signal (trigger) shown in (a),
As shown in (c), the strobe set to 4 ms emits light. The shutter control signal is 1 m as shown in (d).
It is sent after a delay of s and drives the shutter of the camera set to 2 ms as shown in (e). The actual change in strobe light emission intensity is shown in (f).

Both video cameras store the images in their respective fields of view almost at the same time in response to the shutter signal for capturing the images. As described above, the image pickup ranges of the video cameras are partially overlapped with each other, that is, partially overlapped in the width direction of the vehicle travel path. In other words, both video cameras share the range in which the vehicle as the image processing target may exist and capture the image and store the image.

Although it is possible to combine the two images including the license plate portion by image processing,
The processing for that increases and the processing speed decreases. In order to eliminate this, reduce the image processing, and maintain the high-speed operation of the apparatus, in the embodiment, the imaging range of each video camera overlaps the range in which the license plate, which is the main target of the image processing, falls. The license plate is now fully included in the image. An example of each stored image is shown schematically in FIG. An image R ′ 1 is obtained from the video camera 11 and an image R ′ 2 is obtained from the video camera 12.

In this way, either video camera including the license plate can capture the foreground (or the background) of the target vehicle at a high resolution (in a zoomed-up state), and the license plate portion is sufficient. Pixels are associated. In the example of FIG. 4, the video camera 12 obtains an image completely including the license plate portion.

Thereafter, the image processing section 20 sends a camera control signal (S2) for instructing each video camera to output a video signal corresponding to the stored image. As a result, a video signal is input from each video camera to the camera switching unit 21. The image processing unit 20 controls the camera switching unit 21 to selectively output one of the video signals (stored image) from each video camera.

That is, the image processing section 20 first outputs the video signal from the video camera 11 to the image input section 22. Then, each subsequent unit is controlled to perform a series of image processing on this image at least until the license plate portion is extracted. As a result of this processing, if the license plate portion is properly extracted, a predetermined vehicle registration number recognition processing is subsequently performed based on this, and a desired recognition result can be obtained (processing completion).

However, when the license plate cannot be properly obtained even after the license plate extraction process,
The image processing unit 20 controls the camera switching unit 21 again to process the next image, and outputs the video signal (stored image) from the remaining video camera 12 to the image input unit 22. Then, a series of image processing up to license plate extraction is performed on this image in the same manner as described above, and then a predetermined vehicle registration number recognition processing is performed based on the extracted license plate to obtain a desired result. It will be.

The time required for image input is 1 field time (1/60 second) or 1 frame time (1/30 second).
Therefore, the images obtained by dividing the field of view at the same time can be sequentially input with a delay of at most one frame, and the processing time of the image processing apparatus is not sacrificed.

Although it is conceivable to use three or more video cameras, which is larger than that in the embodiment, in such a case,
If the license plate portion cannot be obtained from the image from a certain video camera, the image processing unit 20 sequentially switches a plurality of video cameras until the license plate portion is properly extracted, and inputs the video signal to the license plate portion. The extraction process may be repeated.

FIG. 5 is a block diagram showing another embodiment of the second invention device of the present application. In the device of FIG. 5, the camera switching unit is provided on the video camera side. That is, the video signals from the video camera 11 'and the video camera 12' are selected by the control of the image processing unit 20 'at the camera installation portion, and the selected video signal (stored image) is the image processing unit 20'.
'Is input to the image input unit 22. Only the difference in this point, the operation of the other parts of the illustrated apparatus and each part is almost the same as the case of the above-mentioned embodiment, and therefore the description thereof will be omitted.

Although the second invention of the present application has been mainly described above by taking the case of recognizing the vehicle registration number as an example, the technical contents described above are not limited to the vehicle registration number recognition, and generally, a plurality of objects are processed in order to process a wide imaging target range. The image processing apparatus can be widely applied to an apparatus that divides a processing target area by using a video camera and performs image processing, and can configure an image processing apparatus that significantly reduces the load on the image processing unit side. In this case, the image processing result for determining whether or not to further perform the processing by switching the stored image of each video camera in accordance with the content of each image processing of the image processing unit (in the embodiment, license plate extraction ) Are different from each other and will be appropriately selected.

The image processing apparatus of the first invention of the present application (included in the second invention) is intended for various image processing under the above conditions. That is, the image processing apparatus of the first invention of the present application has a plurality of video cameras and an image processing unit to which these video cameras are connected, and the plurality of video cameras correspond to one screen by an external shutter signal. Image is stored and the stored image is output in response to a selection signal from the outside. The image processing unit sends a shutter signal to each video camera at substantially the same timing, and then, from the video camera, alternatively. The image processing apparatus sequentially performs a series of operations of inputting a stored image and performing image processing on each video camera until a predetermined image processing result is obtained. In this case, a shutter signal (trigger pulse) for image pickup is generated at a predetermined cycle by an oscillator or a timing circuit, or a trigger device suitable for image processing contents is used to store each image substantially at the same time.

Even in the case of a plurality of video cameras, that is, a device for processing a plurality of images, if high-speed processing is not required because the simultaneity of the images is not required, for example, for the purpose of monitoring each installation point, It suffices to simply switch the video signals from each camera and process them in order, and it is not necessary to take the configuration of the present application. The present application is effective only when simultaneity of images and high-speed processing are required.

[0040]

As described in detail above, the image processing apparatus of the first invention of the present application has a plurality of video cameras and an image processing unit to which these video cameras are connected, and the plurality of video cameras are external to each other. An image for one screen is stored by a shutter signal from the device and the stored image is output by a selection signal from the outside. The image processing unit sends the shutter signal to each video camera at substantially the same timing, Since a series of operations for selectively inputting a stored image from a video camera and performing image processing are sequentially performed for each video camera until a predetermined image processing result is obtained, compared to conventional devices. The configuration is simple and can be provided at low cost.

Further, the image processing apparatus of the second invention of the present application includes a plurality of video cameras for obtaining a foreground or a rear view of a traveling vehicle by overlapping and sharing images of the traveling path of the vehicle to obtain images.
An image processing unit that recognizes a vehicle registration number from the output of these video cameras, and the plurality of video cameras stores one screen image by a shutter signal from the image processing unit and stores the image from one image processing unit. The stored image is output to the image processing unit in response to a selection signal. The image processing unit sends a shutter signal to each video camera at substantially the same timing, and then outputs the stored image from each video camera to the camera. A series of operations for image processing of the image selectively input through the switching unit until the license plate is extracted is sequentially performed for each video camera until the license plate is properly extracted. Once extracted, the vehicle registration number is recognized, so
In particular, it is suitable for vehicle registration number recognition, and has the same effect as the first invention, that is, the structure is simpler and less expensive than the conventional device of this type.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a vehicle registration number recognition device which is an embodiment of an image processing device according to the present invention.

FIG. 2 is a plan view illustrating an arrangement of a vehicle registration number recognition device that is an embodiment of the image processing device of both inventions of the present application.

FIG. 3 is a diagram illustrating an operation timing of the apparatus according to the embodiment.

FIG. 4 is a diagram showing an example of a captured image according to the embodiment apparatus.

FIG. 5 is a block diagram showing the configuration of a vehicle registration number recognition device that is another embodiment of the image processing device according to the invention of the present application.

[Explanation of symbols]

 11, 12 (11 ', 12') ... video camera, 20 (20 ') ... image processing unit, 21 ... camera switching unit.

Claims (2)

[Claims] 1. A plurality of video cameras (11, 12) and an image processing section (2) to which these video cameras (11, 12) are connected.
0) and the plurality of video cameras (11, 12) store an image for one screen by a shutter signal from the outside, and output the stored image by a signal from the outside. The image processing unit (20) sends a shutter signal to each video camera (11, 12) at substantially the same timing, and then selectively inputs a stored image from each video camera. An image processing apparatus for sequentially performing a series of image processing operations on each video camera until a predetermined image processing result is obtained. 2. A plurality of video cameras (11, 12) for obtaining a foreground or a foreground of a traveling vehicle by overlapping and sharing images of a vehicle traveling path in the width direction, and these video cameras (11, 12). )
And an image processing unit (20) for recognizing a vehicle registration number from the output of the image processing unit, wherein the plurality of video cameras (11, 12) are the image processing units.
An image for one screen is stored by a shutter signal from (20), and the stored image is output to the image processing unit (20) by a signal from the image processing unit (20). 20) is to output the shutter signal to each video camera at almost the same timing, and then output the stored image from each video camera, and selectively input the image through the camera switching unit (21) to the license plate. A series of operations for image processing until part extraction is sequentially performed for each video camera (11, 12) until the license plate part is appropriately extracted, and after the license plate part is extracted, vehicle registration number recognition is performed. Image processing device.