JP2005310027A - Vehicle traffic investigation system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a vehicle traffic investigation system which can automatically identify vehicle types from vehicle numberplates at high accuracy. <P>SOLUTION: The vehicle traffic investigation system calculates a vehicle width from a vehicle image for every vehicle to be investigated, and determines a vehicle classification number, which is acquired from an image of a classification number that indicates type and usage for vehicles based on a numberplate image acquired from the vicinity of the center of the vehicle width by pattern comparison using font data corresponding to figures 0 to 9. In addition, the system acquires a vehicle type code, which is determined from the vehicle classification number and a numberplate size corresponding to the vehicle width with reference to the large, middle and small numberplate sizes specified by the Road Transportation Vehicle Low. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a vehicle traffic survey system, and more particularly to a vehicle traffic survey system having a function of automatically identifying a vehicle type from a license plate of a vehicle.

  Vehicle traffic surveys are carried out for each prefecture and municipality for city improvement plans and road plans, and once a few years, they are conducted nationwide under the control of the Ministry of Land, Infrastructure, Transport and Tourism.

  Conventionally, vehicle traffic surveys have been conducted by counting the number of vehicles that pass through a predetermined part of the road visually by an investigator, and determining the vehicle type from the license plate. However, there are many labor costs and errors. Had a problem.

  Therefore, a passing vehicle is photographed by a photographing device installed at a predetermined place to acquire image data, and the image data is transmitted to the processing center, and the license plate in the image data is visually determined while viewing the display screen at the processing center. However, the above problem has not been solved fundamentally.

  By the way, the following Patent Documents 1 and 2 are listed as techniques for automatically performing a traffic survey by photographing a passing vehicle with a photographing device installed at a predetermined location to acquire image data and analyzing the image data. is there.

The traffic investigation system described in Patent Document 1 below captures a passing vehicle by using photographing devices installed at a plurality of points, acquires image data at each point, cuts out a license plate image in each image data, and obtains license plate information. Is stored together with the time, and the passing time is obtained from the difference between the reading times at two different shooting points, and when the passing time is equal to or shorter than the predetermined time, the vehicle is simply passed between the two points. Processing that is considered.
Japanese Patent Application Laid-Open No. 2002-175587

In addition, the traffic volume measuring device described in Patent Document 2 described below determines the vehicle type of a vehicle entering the service area by using a license plate recognition device installed on the entrance side travel path of the service area, and the exit side travel path of the service area. Vehicles leaving the service area are detected by a detector installed in the vehicle, and based on these pieces of information, the vehicles in the service area are classified for each vehicle type and a traffic volume survey is performed.
JP 2002-24988 A

  On the license plate, letters (such as “Shinagawa”) that indicate the land transportation branch office or automobile inspection registration office of the headquarters in use, a classification number (such as “330”) according to the type and use of the vehicle, and business, private, and rental cars A character (such as “sa”) indicating a type such as a business name and a 4-digit series designation number (such as “12-34”) are displayed.

  In addition, there are three types of license plates: large plate size (length 220 mm × width 440 mm), medium plate size (length 165 mm × width 330 mm), and small plate size (length 125 mm × width 230 mm).

  Large plate size license plates start with a large truck with a classification number starting with "1", a large passenger car with a classification number starting with "2" (with a passenger capacity of 30 or more), and "8" Mounted on special purpose vehicles with classification numbers (applicable to large vehicles).

  Medium-size license plates are classified as ordinary trucks with classification numbers starting with "1", large passenger cars with classification numbers starting with "2" (with 11 to 29 passengers), classifications starting with "3" Ordinary passenger cars with numbers, small lorries with classification numbers starting with "4" and "6", small passenger cars with classification numbers starting with "5" and "7", special classification numbers starting with "8" Used in automobiles (applicable to ordinary and small cars), large special cars with classification numbers starting with “9”.

  Medium plate size license plates can also be used for light trucks with classification numbers beginning with “4”, light duty vehicles with classification numbers beginning with “5”, and special purpose vehicles with classification numbers beginning with “8” (light Also applicable to automobiles).

  The plate size license plate is a two-wheeled light vehicle with classification number “1” or “2”, a three-wheeled light vehicle with classification number “3” or “33” (registered in the first year by 1974), Light trucks with classification number “6” or “66” (registered in the first year by 1974), mini passenger cars with classification number “8” or “88” (registered in the first year by 1974) Special-purpose vehicles with classification number “0” or “00” (those that fall under the first year registration by 1974) and motorcycles that do not have a classification number It is attached.

  Therefore, the vehicle type of the vehicle can be determined by the classification number and size of the license plate of the vehicle.

  However, in the prior art described in Patent Documents 1 and 2, all displays must be acquired as text data from the license plate image in the image data and collated, resulting in a huge amount of data and a long time for analysis. Cost.

  Further, these conventional techniques are not based on the technical idea of determining the vehicle type of the vehicle based on the classification number and size of the license plate of the vehicle.

  Further, in these conventional techniques, a photographing device for acquiring a vehicle image is fixed to a structure such as a pedestrian bridge over the road, or installed on the side of the road, and the vehicle is photographed from the front. However, since the number plate of the two-wheeled vehicle is attached only to the rear portion, the number plate image of the two-wheeled vehicle cannot be acquired according to this method.

  SUMMARY OF THE INVENTION Accordingly, an object of the present invention is to provide a vehicle traffic survey system that eliminates the above-mentioned disadvantages of the prior art and can automatically identify a vehicle type from a vehicle license plate with high accuracy.

  In order to achieve this object, the present invention according to claim 1 is obtained by a vehicle width calculation means for calculating a vehicle width from a vehicle image for each vehicle to be investigated, and the vehicle width calculation means in a vehicle traffic investigation system. A license plate image obtaining means for obtaining an image of a license plate near the center of the vehicle width, and a vehicle classification for obtaining an image of a classification number according to the type and use of the vehicle from the license plate image obtained by the license plate image obtaining means. Vehicle classification number acquisition means, and vehicle classification number acquisition means for acquiring a vehicle classification number by pattern matching using font data corresponding to the numbers 0 to 9, based on the vehicle classification number image by the vehicle classification number image acquisition means It is characterized by providing.

  The present invention according to claim 2 is the vehicle traffic survey system according to claim 1, wherein the number plates of large plates, medium plates, and small plates defined by the Road Transport Vehicle Law are determined by the vehicle width calculated by the vehicle width calculating means. Number plate size determination means for determining which size the license plate of the vehicle corresponds to among the sizes, and a vehicle type code determined by the combination of the license plate size and the vehicle classification number prepared in advance by the user The vehicle type code list storage means for storing the correspondence table, the license plate size by the license plate size determination means and the vehicle classification number by the vehicle classification number acquisition means are collated with the list read from the vehicle type code list storage means. And a vehicle type code discriminating means for determining the corresponding vehicle type code. That.

  According to a third aspect of the present invention, in the vehicle traffic survey system according to the first or second aspect, the vehicle width calculation means can calculate the vehicle width of the contour line of the vehicle from the vehicle image for each vehicle to be surveyed. It is characterized in that end point coordinates connecting such specific contour lines are acquired, and the vehicle width is calculated by correcting an error due to the camera angle at the time of shooting from the end point coordinates.

  According to a fourth aspect of the present invention, in the vehicle traffic survey system according to any one of the first to third aspects, the license plate image acquisition means includes a large plate, a middle plate, and a small plate defined by the Road Transport Vehicle Law. The number plate size for each license plate size is stored for each license plate size, and the license plate size determined by the number plate size determination unit is determined for each license plate size from the vehicle image for each vehicle to be investigated. A rectangular image is acquired with respect to the acquisition surface, centering on the bisection point of the vehicle width obtained by the vehicle width calculation means, for the area corresponding to the dimensions obtained by comparing with the list read from the dimension list storage means. The image acquisition means for the area for the license plate size and the image acquisition means for the area for the license plate size License plate image tilt correction means for correcting a license plate image having a tilt corresponding to the angle of the camera at the time so that it becomes a rectangular image as if taken from the front. .

  According to a fifth aspect of the present invention, in the vehicle traffic survey system according to any one of the first to fourth aspects, still image data acquired at predetermined time intervals from moving image data of a camera that images the roadway. Image input means for storing the input image data together with the photographing time, vehicle presence determination means for determining whether or not there is a vehicle in the input image data, and vehicle presence determination means Vehicle identification means for acquiring a cut-out image for each identified vehicle based on input image data in which a vehicle may exist and storing it as vehicle data together with a vehicle ID that is a unique number of each vehicle; It is further provided with the feature.

  According to a sixth aspect of the present invention, in the vehicle traffic survey system according to the fifth aspect, the photographing means for providing the moving image data that is the basis of the still image data acquired by the image input means at regular time intervals, A motorcycle having a license plate only at the rear of the vehicle can be photographed by installing a photographing device on the shoulder of the roadway and photographing a passing vehicle in the traveling direction of the vehicle.

  According to a seventh aspect of the present invention, in the vehicle traffic survey system according to the fifth or sixth aspect, the vehicle presence determination means may include a vehicle for the first input image having no previous input image. For an input image with a previous input image that is determined to have the following characteristics, the tracking vehicle list that stores the vehicle ID of the vehicle recognized in the previous input image is referred to, and the tracking vehicle If there is a vehicle, the vehicle may be present, and if there is no tracked vehicle, the current input image is compared with the input image immediately before the vehicle does not exist. If there is a difference, it is determined that there is a possibility that a vehicle exists in the current input image, and if there is no difference, it is determined that there is no vehicle.

  According to an eighth aspect of the present invention, in the vehicle traffic survey system according to any one of the fifth to seventh aspects, the vehicle identification means binarizes the input image based on the input image data. A contour emphasizing unit for emphasizing the contour of the subject, a wire frame forming unit for recognizing the subject as a wire frame by following a contour line by the contour emphasizing unit, and an object recognized by the wire frame forming unit. The vehicle shape identifying means for recognizing the subject having the shape of the vehicle and having an appropriate size as the vehicle as the vehicle, and the vehicle ID of the vehicle recognized for each input image data are stored. For each vehicle identified by the tracked vehicle list creating unit and the vehicle form identifying unit, the tracked vehicle list of the previous input image by the tracked vehicle list creating unit is retained. If the vehicle can be identified as the same vehicle by collating with the vehicle data of each vehicle ID, the vehicle is determined as a tracking vehicle. Vehicle tracking means for storing.

  The present invention according to claim 9 is the vehicle traffic survey system according to claim 8, wherein the vehicle form identification means forms an end point pattern list for forming a wire frame for each shape of a freight car, a normal passenger car, a large passenger car, etc. A shape pattern list storage means, and a shape corresponding to the shape of the vehicle to be identified is determined from the list read from the shape pattern list storage means, and the determination result is used as shape data of the vehicle data. It further comprises a vehicle type code determination unit based on shape data, wherein a corresponding vehicle type code is determined based on a combination of the shape data and the plate size obtained by the license plate size determination unit.

  According to a tenth aspect of the present invention, in the vehicle traffic survey system according to any one of the second to ninth aspects, vehicle data recognized after the previous counting time is extracted at regular time intervals. It further comprises a number counting means for counting the number of vehicles for each vehicle type code and storing it as number counting data.

  The present invention according to claim 11 is the vehicle traffic survey system according to claim 10, further comprising communication means for transmitting the total number data to a processing center that collectively counts the total number data of each survey point. It is characterized by.

  ADVANTAGE OF THE INVENTION According to this invention, the disadvantage fault of a prior art can be eliminated, and the vehicle traffic investigation system which can identify a vehicle type automatically from a vehicle license plate with high precision can be provided.

(Outline of system configuration)
As shown in FIG. 1, the system is photographed based on a photographing device 104 that is a camera that is installed on the shoulder of a road in the traveling direction of a roadway and photographs a passing vehicle, and image data photographed by the photographing device 104. The vehicle classification recognition processing unit 100 for recognizing the plate size, classification number, and vehicle type for each passing vehicle, and the number counting unit for counting the number of vehicle types based on the vehicle type data for each vehicle of the vehicle classification recognition processing unit 100 102 and a communication device 106 which is a portable communication terminal for transmitting the total data of the number totalization unit 102 to the processing center 108.

(Detailed description of the system)
The vehicle classification recognition processing unit 100 including the vehicle type classification processing unit 206, which is the core of this system, will be described in detail below with reference to FIGS. The vehicle type classification processing unit 206 is a unit that discriminates the classification number of the vehicle license plate and the vehicle type code.

(S1)
In step S <b> 1, the image input unit 200 acquires still image data at regular time intervals from the moving image data storage device 220 that stores moving image data captured by a camera that is fixedly shooting the roadway. Are created together with the shooting time as input image data (see FIG. 4) and stored in the input image data storage device 222.

(S2)
In step S <b> 2, the vehicle presence determination unit 202 determines whether there is a vehicle or a vehicle may exist, based on the input image data read from the input image data storage device 222. Since the processing unit is already defined here, details will be described later.

(S3)
In step S3, the vehicle identification unit 204 assigns a vehicle ID to each vehicle included in the image area shown in FIG. 5 based on the input image data read from the input image data storage device 222. Track and manage until it disappears from the area. Since the processing unit is already defined here, details will be described later.

(S4)
In step S4, for each vehicle recognized in step S3, the vehicle type classification processing unit 206 sequentially determines the plate size, classification number, and vehicle type code of the vehicle. Since the processing unit is already defined here, details will be described later.

(Details of S2)
Next, the process of the vehicle presence determination part 202 of step S2 is explained in full detail below with reference to FIG.

(S2-1)
In step S2-1, since it is the first input image at the start of the traffic survey, the determination unit that determines whether or not there is an immediately preceding input image does not include the immediately preceding input image. It is determined that there is a possibility that it exists, and the process is continued in the vehicle identification unit 204 in the next step S3. If there is an input image immediately before that, the tracking vehicle presence / absence determination unit in step S2-2 is determined. Processing is continued.

(S2-2)
In step S2-2, the tracking vehicle presence / absence determination unit refers to the tracking vehicle list (FIG. 7) in the input image immediately before the vehicle classification recognition processing unit 100 holds. It is determined that there is no, and the processing is continued in the comparison operation unit with the input image immediately before the next step S2-3. If there is one or more, there is a tracking vehicle, and there is also a vehicle in the current input image. It is determined that there is a possibility, and the process is continued in the vehicle identification unit 204 in the next step S3.

(S2-3)
In step S2-3, the comparison calculation unit with the previous input image compares the current input image with the previous input image without the tracking vehicle, and there is a possibility that a vehicle exists if there is a difference. Otherwise, it is determined that there is no vehicle.

  More specifically, the comparison operation is performed by calculating a difference portion by exclusive OR (XOR) for each pixel having the same coordinates between two images of the current input image and the previous input image.

  However, here, instead of recognizing the vehicles individually, to identify whether or not the vehicle exists as a whole image, if an object such as a bird or a shadow appears in the image area, it will be As a result of the comparison calculation, there is a difference. As a result, it is determined that there is a vehicle even though there is no vehicle. However, even if this determination result is issued for reasons other than the vehicle, there is no problem because only the vehicle is identified by the vehicle identification unit 204 in the next step S3.

  In addition, regarding the difference of the comparison calculation between the two images, it is possible to improve the determination accuracy by providing a threshold value for the size of the difference value and the size of the area occupied by the difference.

(Details of S3)
Next, the process of the vehicle identification part 204 of step S3 is explained in full detail below with reference to FIG.

(S3-1)
In step S3-1, the image contour emphasis processing unit generates an image in which the contour of the object is emphasized by binarizing the input image.

(S3-2)
In step S3-2, the subject is recognized as a wire frame by tracing the contour line based on the contour-enhanced image generated in step S3-1.

(S3-3)
In step S3-3, the vehicle form identification process extracts the vehicle shape from the wire frame generated in step S3-2.

  In brief, the vehicle having the shape of a vehicle has the shape of a rectangular parallelepiped viewed from the obliquely rear viewpoint shown in FIG. A wire frame of a subject that is similar to this shape and can be determined to have an appropriate size as a vehicle can be determined as a vehicle.

  More specifically, in the vehicle form identification process, in order to obtain the vehicle shape data necessary for the alternative process when the normal vehicle type code determination by the size of the license plate and the classification number is not possible, The following processing is performed. As illustrated in FIG. 10, the shape of the vehicle includes a list of a plurality of end point coordinates connecting a vehicle shape pattern such as a freight car, a normal passenger car, and a large passenger car (bus) and an outline characterizing each vehicle shape. A similar shape pattern is discriminated from a combination of the above and a vehicle having an appropriate size is selected. Note that the end point coordinate list is the one in which the end points are arranged in the order of the list and connected to make it possible to recall the vehicle shape from the wire frame image. Not) After determining the shape pattern of the vehicle, the shape pattern number and a coordinate list (also referred to as wire frame data) arranged in the order of the end point list corresponding to the vehicle shape pattern are temporarily stored.

  The processing image from the above steps S3-1 to S3-3 is shown in FIG.

(S3-4)
In step S3-4, the vehicle identification number assignment / identification number assigned vehicle tracking unit compares each vehicle identified in step S3-3 with each vehicle in the tracked vehicle list in the immediately preceding input image. Whether the vehicle is a new vehicle or a tracking vehicle is determined from the relationship between the vehicle position that can be calculated from the vehicle shape pattern or the wire frame data of the individual vehicle data and the traveling direction of the roadway.

  When it is determined that the vehicle is a new vehicle, vehicle data (see FIG. 12) in which the wire frame data of the vehicle and the shape pattern number are written as the shape data is created and stored in the vehicle data storage device 224. Or when it determines with a tracking vehicle, the present wire frame data is added in applicable vehicle data.

(Details of S4)
Subsequently, the process of the vehicle type classification processing unit 206 in step S4 will be described in detail below with reference to FIGS. A vehicle type classification process is performed for each vehicle recognized in step S3. However, FIG. 13 shows only a functional unit and a data storage device that are the main points of the processing unit.

(S4-1)
In step S4-1, the vehicle type code determined confirmation unit determines whether the vehicle type code has been determined based on whether or not a value is input to the vehicle type code of the target vehicle based on the vehicle data created in step S3. Determine whether or not. If it is determined, the process of step S4 is terminated, otherwise the next step is continued.

(S4-2)
In step S4-2, the license plate recognition range determination unit determines whether or not the vehicle is in the license plate identification range by using the wire frame data in the vehicle data. As a result, if it is out of the range, the process of step S4 is terminated, and if it is within the range, the next step is continued. The license plate identification range is determined in advance by the user of the present system as a range that does not hinder license plate recognition in the image area shown in FIG.

(S4-3)
In step S4-3, the vehicle size measuring unit calculates the vehicle width and the like from the wire frame data in the vehicle data according to the equation of FIG. Here, the function of calculating the vehicle width corresponds to the vehicle width calculation unit 300 of FIG.

  The bit length coefficient is an index value for converting the length of each pixel in the image into an actual measurement value, and is related to the distance between the camera and the subject.

(S4-4)
In step S4-4, the license plate size determination unit 302 determines, based on the vehicle width of the vehicle size acquired in step S4-3, out of the license plate sizes of large plates, medium plates, and small plates determined by the Road Transport Vehicle Law. Then, the size of the license plate of the vehicle is determined, and the determination result is written in the vehicle data.

  The license plate size is determined by the vehicle width when the vehicle width is greater than 1.7 m, the large plate size, when the vehicle width is greater than 1.48 m and less than 1.7 m, the middle plate size and the vehicle width are 1 When it is less than 48m, it is designed to be a small plate size.

(S4-5)
In step S4-5, the license plate image acquisition unit 304 acquires a license plate image near the center of the vehicle width of the vehicle size acquired in step S4-3. Since the processing unit is already defined here, details will be described later.

(S4-6)
In step S4-6, the vehicle classification number acquisition unit 310 from the license plate image changes the numbers 0 to 9 based on the vehicle classification number image acquired from the number plate image by the number plate image acquisition unit 304 in step S4-5. The classification number of the target vehicle is acquired by pattern matching using the corresponding font data, and the acquisition result is written in the vehicle data. Since the processing unit is already defined here, details will be described later.

(S4-7)
In step S4-7, the vehicle type code determination unit 312 generates a list of vehicle type codes (see FIG. 15), which is determined by the combination of the license plate size and the vehicle classification number, which is created in advance by the user. The vehicle type code corresponding to the combination of the license plate size and the vehicle classification number in the vehicle data is acquired from 322, and the acquired vehicle type code is written in the vehicle data. However, there is an exception that the vehicle type code cannot be determined only by the license plate size and the vehicle classification number in FIG. 15, but in that case, the vehicle type code is determined by combining the determination in the next step S4-8. It is also possible.

(S4-8)
In step S4-8, the vehicle type code determination unit based on shape data cannot acquire a license plate image in step S4-5 because the license plate image acquisition unit 304 is congested on the roadway and vehicles are connected in a daisy chain. If the vehicle classification number acquisition unit 310 from the license plate image cannot acquire the vehicle classification number for some reason in step S4-6, the vehicle type code determination unit 312 in step S4-7 does not function. An auxiliary vehicle type code is discriminated.

  Specifically, the vehicle type code determination unit based on the shape data refers to a list of vehicle type codes (see FIG. 16) prepared by the user and determined by the combination of the license plate size and the shape data. The vehicle type code corresponding to the combination of the shape data of the vehicle data and the license plate size is acquired, and the acquired vehicle type code is written in the vehicle data. However, the shape data and shape examples in FIG. 16 do not cover all shapes, but are shown as examples only.

  The vehicle type code determination unit based on this shape data functions effectively as a recovery of the vehicle type code determination process for special cases such as traffic jams and modified vehicles, although the accuracy is somewhat reduced.

(Details of S4-5)
Next, the processing of the license plate image acquisition unit 304 in step S4-5 will be described in detail below with reference to FIG.

(S4-5-1, S4-5-2)
First, the image acquisition means for the area corresponding to the license plate size stores a dimension list for each license plate size (FIG. 18, FIG. 19) is read from the dimension list storage device 320 for each license plate size (step S4-5-1), and the dimension corresponding to the license plate size of the vehicle data is obtained by referring to the dimension list. The image of the license plate area is cut out and acquired for the area according to the dimension, with the bisection point of the vehicle width of the vehicle data as the reference (center) and the shape as the longer side of the cut-out surface (step S4- 5-2). It should be noted that the bisector of the vehicle width of the vehicle data is accurately calculated from the midpoint coordinates of the end points corresponding to the end points a and b in FIG. 9 of the wire frame data of the vehicle.

(S4-5-3)
In step S4-5-2, the license plate image inclination correcting unit corrects the image of the license plate area acquired in step S4-5-1 by an amount corresponding to the inclination of the camera angle, as if it was photographed from the front. Make it a rectangular image.

(Details of S4-6)
Subsequently, the processing of the vehicle classification number acquisition unit 310 from the license plate image in step S4-6 will be described in detail below with reference to FIG.

(S4-6-1, S4-6-2)
First, in order to acquire an image of the classification number according to the type and use of the vehicle based on the license plate image that has been tilt-corrected by the license plate image acquisition unit 304 in step 4-5, an area list of classification numbers for each license plate size (Refer to FIGS. 18 and 19) is read out (step S4-6-1), the area corresponding to the license plate size of the vehicle data is obtained by referring to the area list, and the upper left (X, Y) Is used to cut out and obtain the classification number images corresponding to the rectangular area having the width W and the height H (step S4-6-2). Note that the vehicle classification number image acquisition unit 308 of FIG. 13 corresponds to step S4-6-2.

(S4-6-3)
In step S4-6-3, the classification number acquisition unit from the classification number image collates the classification number image acquired in step S4-6-2 with the license plate font pattern. Pattern matching is sequentially performed using license plate font patterns from 0 to 9 which are vehicle classification codes, and the matched value is written in the vehicle data as a classification number. (Fig. 21)
The above is the vehicle classification recognition processing unit 100 which is the core of the present system. In this embodiment, a processing unit for calculating the traffic volume by vehicle type, which is a normal survey item in the traffic volume survey, is also provided.

  The number counting unit 102 extracts vehicle data recognized after the previous counting time from the vehicle data storage device 224 storing the vehicle data created by the vehicle classification recognition processing unit 100 at regular time intervals, and the extracted data is extracted. From the vehicle data, the number of vehicles for each vehicle type code is totaled and stored as unit count data (see FIG. 22).

  Furthermore, by providing the communication device 106, it is possible to transmit the total number data to the processing center 108 that collectively counts the total number data of each survey point, and it is possible to cope with a wide-area traffic survey on a nationwide scale. Yes.

It is a conceptual diagram of this system. It is a block block diagram of a vehicle classification recognition processing system. It is a figure which shows the processing flow of this system. It is a data structure figure of input image data. It is a figure which shows the effect | action of an imaging device roughly. It is a figure which shows the processing flow of a vehicle presence determination part. It is a figure which shows a tracking vehicle list. It is a figure which shows the processing flow of a vehicle identification part. It is a figure which shows the original model and vehicle size measurement model of a vehicle shape. It is a figure which shows a shape pattern list. It is a figure which shows the process image of a vehicle identification part. It is a data structure figure of vehicle data. It is a block block diagram of a vehicle classification processing part. It is a figure which shows the processing flow of a vehicle type classification | category process part. It is a figure which shows a vehicle type code list. It is a figure which shows the vehicle type code list by shape data. It is a figure which shows the processing flow of a license plate image acquisition part. It is a figure which shows the area | region list | wrist of the dimension for every license plate size, and a classification number. It is a figure which shows the relationship between the dimension for every license plate size, and a classification number. It is a processing flowchart of the vehicle classification number acquisition part from a license plate image. It is a figure which shows the processing image of the pattern collation of a classification number. It is a figure which shows a number total data list.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 Vehicle classification recognition process part 102 Number totalization part 104 Image pick-up device 106 Communication apparatus 108 Processing center 200 Image input part 202 Vehicle presence determination part 204 Vehicle identification part 206 Vehicle type classification process part 220 Moving image data storage apparatus 222 Input image data (still image) ) Storage device 224 Vehicle data storage device 226 Number counting data storage device 300 Vehicle width calculation unit 302 License plate size determination unit 304 License plate image acquisition unit 308 Vehicle classification number image acquisition unit 310 Vehicle classification number acquisition unit 312 Vehicle type code determination unit 320 Dimension list storage device for each license plate size 322 Vehicle type code list storage device

Claims (11)

Vehicle width calculation means for calculating the vehicle width from the vehicle image for each vehicle to be investigated;
A license plate image acquisition means for acquiring an image of a license plate near the center of the vehicle width obtained by the vehicle width calculation means;
Vehicle classification number image acquisition means for acquiring an image of a classification number according to the type and use of the vehicle from the license plate image by the license plate image acquisition means;
Vehicle classification number acquisition means for acquiring a vehicle classification number by pattern matching using font data corresponding to the numbers 0 to 9, based on the vehicle classification number image by the vehicle classification number image acquisition means;
A vehicle traffic survey system comprising: According to the vehicle width calculated by the vehicle width calculating means, it is determined which size the license plate of the vehicle corresponds to among the license plate sizes of the large plate, the middle plate, and the small plate defined by the Road Transport Vehicle Law. Plate size determination means;
Vehicle type code list storage means for storing a correspondence table of vehicle type codes determined by a combination of a license plate size and a vehicle classification number, which has been created in advance by a user;
A vehicle type code determining unit for determining a corresponding vehicle type code by comparing the license plate size by the license plate size determining unit and the vehicle classification number by the vehicle classification number acquiring unit with a list read from the vehicle type code list storage unit;
The vehicle traffic survey system according to claim 1, further comprising:   The vehicle width calculation means obtains end point coordinates connecting a specific contour line capable of calculating the vehicle width from the vehicle image for each vehicle to be investigated, and from the end point coordinates according to the camera angle at the time of shooting. The vehicle traffic investigation system according to claim 1 or 2, wherein the vehicle width is calculated by correcting an error. The license plate image acquisition means includes
Dimension list storage means for each license plate size, storing dimensions for each license plate size of the large plate, middle plate, and small plate defined by the Road Transport Vehicle Law;
From the vehicle image for each vehicle to be investigated, the vehicle is only an area corresponding to the dimension obtained by comparing the license plate size determined by the license plate size determination means with the list read from the dimension list storage means for each license plate size. Centering on the bisection point of the vehicle width obtained by the width calculation means, an image acquisition means for acquiring a rectangular image with respect to the acquisition surface, an area for the license plate size,
Correcting the license plate image obtained by the image acquisition means for the area corresponding to the license plate size and having an inclination corresponding to the angle of the camera at the time of shooting so that it becomes a rectangular image as if taken from the front, License plate image tilt correction means,
The vehicle traffic investigation system according to any one of claims 1 to 3, further comprising: Image input means for storing still image data acquired at predetermined time intervals from moving image data of a camera that images the roadway as input image data together with the shooting time;
Vehicle presence determination means for determining whether or not there is a vehicle in the input image data or a vehicle may exist;
A vehicle that acquires a cut-out image for each identified vehicle based on input image data in which the vehicle may be present by the vehicle presence determination means, and stores it as vehicle data together with a vehicle ID that is a unique number of each vehicle An identification means;
The vehicle traffic investigation system according to claim 1, further comprising:   The photographing means for providing the moving image data that is the basis of the still image data acquired by the image input means at regular time intervals sets the photographing device on the shoulder of the roadway and photographs the passing vehicle in the vehicle traveling direction. The vehicle traffic investigation system according to claim 5, wherein a motorcycle having a license plate only at the rear of the vehicle can be photographed. The vehicle presence determination means
For the first input image that has no previous input image, it is determined that there may be a vehicle,
For input images that have a previous input image after that,
With reference to the tracked vehicle list storing the vehicle ID of the vehicle recognized in the immediately preceding input image, if the tracked vehicle exists, it is determined that the vehicle may exist, and the tracked vehicle Is not present, the current input image is compared with the input image immediately before the vehicle does not exist, and if there is a difference, it is determined that there is a possibility that the vehicle exists in the current input image. The vehicle traffic investigation system according to claim 5 or 6, wherein if there is no difference, it is determined that no vehicle exists. The vehicle identification means includes
Contour emphasizing means for enhancing the contour of the subject by binarizing the input image based on the input image data;
Recognizing the subject as a wire frame by following the contour line by the contour emphasizing unit;
Vehicle form identifying means for recognizing a subject having a vehicle shape and having a reasonable size as a vehicle from among the subjects recognized by the wire frame forming means;
Tracking vehicle list creation means for storing the vehicle ID of the vehicle recognized for each input image data;
If each vehicle identified by the vehicle form identifying means can be identified as the same vehicle by collating with the vehicle data of each vehicle ID held in the tracked vehicle list of the immediately preceding input image by the tracked vehicle list creating means, Vehicle tracking means for acquiring the vehicle image as a new vehicle and storing it as vehicle data together with the vehicle ID;
The vehicle traffic investigation system according to any one of claims 5 to 7, further comprising: The vehicle form identification means includes a shape pattern list storage means for storing an end point pattern list for forming a wire frame for each shape of a cargo vehicle, a normal passenger car, a large passenger car, etc., and is read from the shape pattern list storage means In the list, the shape corresponding to the shape of the identification target vehicle is determined, and the determination result is written as shape data of the vehicle data.
9. The vehicle traffic investigation system according to claim 8, further comprising vehicle type code determination means based on shape data for determining a corresponding vehicle type code based on a combination of the shape data and the plate size obtained by the license plate size determination means. .   It further comprises a number counting means for extracting vehicle data recognized after the previous counting time at a certain time interval, totaling the number of vehicles for each vehicle type code, and storing it as number counting data. The vehicle traffic investigation system according to any one of claims 2 to 9.   11. The vehicle traffic survey system according to claim 10, further comprising a communication unit that transmits the total number data to a processing center that collectively counts the total number data at each survey point.

Images