JP7036615B2 - Vehicle recognition device and parking lot management system - Google Patents

Description

The present invention relates to a vehicle recognition device that recognizes a vehicle by a vehicle number or the like and a parking lot management system that manages a parking lot by using the vehicle recognition device.

A technique for identifying a vehicle by reading a vehicle number, that is, for recognizing a vehicle number, has been conventionally known (see, for example, Patent Document 1).

Regarding vehicle number recognition, there is a problem that it becomes unreadable due to the external environment, etc., and in Patent Document 1, similar relationships are listed and a search and determination are performed by expanding the permissible range based on the retained data. We are trying to improve the recognition rate.

However, in recent years, new characters and patterns such as so-called local numbers have increased, and the design of automobile license plates has diversified, which may lower the recognition rate in particular. In response to this, it is conceivable to upgrade the recognition engine applied to vehicle number recognition each time, but the system recognition rate will continue to decline until a new recognition engine is released. It will be.

Japanese Unexamined Patent Publication No. 2016-224773

The present invention has been made in view of the above points, and regarding vehicle recognition, for example, even if a new character, a symbol number, etc. is added, it can be immediately dealt with, and vehicle recognition can improve the recognition rate. It is an object of the present invention to provide a parking lot management system using a device and a vehicle recognition device.

The vehicle recognition device for achieving the above object extracts a feature amount unique to the vehicle including the vehicle number from the image data acquisition unit that acquires the image data of the vehicle and the image data acquired by the image data acquisition unit. It includes a feature amount extraction unit and a vehicle identification unit that identifies a vehicle based on the information extracted by the feature amount extraction unit.

In the above vehicle recognition device, when identifying a vehicle, not only the vehicle number but also the feature amount peculiar to the vehicle is used, so that even if some vehicle numbers cannot be read, for example, the vehicle It is possible to reliably identify the vehicle by the feature amount. That is, even when a new character, a pattern number, or the like is added, it is possible to immediately respond to this by using the feature amount of the vehicle, and it is possible to improve the recognition rate. Further, for example, if the vehicle recognition device is used in parking lot management, it is possible to more reliably determine whether or not to allow the vehicle to pass through when entering or exiting the parking lot.

In a specific aspect of the present invention, the feature amount extraction unit is a vehicle number reading unit that reads a vehicle number as a feature amount to be extracted, and a feature amount related to a vehicle image that is a feature amount other than the vehicle number among the feature amounts to be extracted. It has a feature amount calculation unit for calculating. In this case, the vehicle can be specified by using the vehicle number information and the feature amount related to the vehicle image other than the vehicle number as the feature amount peculiar to the vehicle.

In another aspect of the present invention, in the feature amount calculation unit, the feature amount relating to the vehicle image relates to at least one of the shade of the color of the license plate, the shape of the license plate, and the shape of the vehicle member around the license plate. Includes feature quantities. In this case, such as special license plates such as local license plates, the characteristics of color shades and shape characteristics due to the inclusion of new characters and patterns that are not found in ordinary license plates, or license plates. The vehicle can be identified by capturing the shape characteristics of the vehicle in places other than the above.

In yet another aspect of the present invention, the vehicle identification unit relates to a vehicle image calculated by the feature amount calculation unit when the vehicle cannot be specified only by the vehicle number information that can be read by the vehicle number reading unit. The feature amount information is used to identify the vehicle. In this case, if the vehicle number can be read reliably, the entry / exit process can be performed quickly, and even if the vehicle number cannot be read, the feature amount information related to the vehicle image can be obtained. By using it, you can process entry and exit.

In yet another aspect of the present invention, whether or not the vehicle identification unit is the same as one vehicle on the database by matching the feature amount of the vehicle on the registered database with the feature amount of the specified vehicle. Is determined. In this case, it is possible to quickly and reliably perform entry / exit processing based on whether or not the vehicle is a registered vehicle.

In order to achieve the above object, the parking lot management system according to the present invention manages entry / exit of the parking lot based on any of the above vehicle recognition devices and the feature amount of the vehicle acquired by the vehicle recognition device. It has a participation management department. In this case, by providing any of the above vehicle recognition devices, it is possible to more reliably determine whether or not to allow the vehicle to pass through when entering or exiting the vehicle.

In a specific aspect of the present invention, the vehicle specifying unit registers the specified vehicle feature amount in the database when the vehicle enters, and identifies the vehicle feature amount on the registered database when the vehicle enters. By matching with the feature amount of the vehicle, it is determined whether or not the same vehicle exists, and when the entry / exit management department determines that the same vehicle exists in the database in the vehicle identification unit when the vehicle enters. Permission to participate in. In this case, by using the data at the time of entering the vehicle, it is possible to reliably identify the vehicle at the time of entering the vehicle.

It is a block diagram explaining one configuration example of the parking lot management system including the vehicle recognition device which concerns on 1st Embodiment. It is a block diagram explaining one configuration example of a vehicle recognition device. (A) is a front view conceptually showing a license plate or image data obtained by reading a license plate, (B) is a front view showing an example of a special specification license plate, and (C) is a number. It is an image diagram of an example which extracted the peripheral part of a plate and a license plate. (A) is a diagram showing an example of a list of read vehicle data, and (B) is a diagram showing an example of feature amount data which is information indicating a feature amount related to a vehicle image among vehicle data. Is. It is a flowchart which shows an example of a series of processing at the time of admission to a parking lot. It is a flowchart which shows an example of a series of processing at the time of participation from a parking lot. It is a block diagram about a modification of the parking lot management system provided with a vehicle recognition device. It is a block diagram about another modification of the parking lot management system provided with a vehicle recognition device. It is a block diagram about another modification of the parking lot management system provided with a vehicle recognition device. It is a block diagram explaining one configuration example of the parking lot management system including the vehicle recognition device which concerns on 2nd Embodiment. It is a flowchart which shows an example of a series of processing at the time of admission to a parking lot.

[First Embodiment]
Hereinafter, an example of the parking lot management system according to the first embodiment of the present invention will be described with reference to FIG. 1 and the like. The parking lot management system 500 according to the present embodiment includes a parking lot management device 100, an entrance unit 10 that manages an entrance for entering a vehicle, and an entry unit 30 that manages an exit for entering a vehicle. In the illustrated example, the parking lot management system 500 is a management system that assumes a toll parking lot mainly for vehicles that are temporarily used as a management target. In the figure, arrows A1 and A2 indicate the direction of entry and exit of the vehicle, respectively.

Of the parking lot management system 500, the parking lot management device 100 is a parking lot management device for comprehensively controlling various types of management in the parking lot, and is configured by, for example, a management server, a PC, or the like to perform various arithmetic processes. In addition to the CPU, it is composed of a memory for storing various data and programs. Therefore, for example, the parking lot management device 100 holds various data and stores newly created data in a predetermined area. In addition, an image of acquiring image data from the entrance section 10 and the entry section 30. A data acquisition unit 50, a feature amount extraction unit 60 for extracting vehicle characteristics, a vehicle identification unit 70 for specifying a vehicle, and a gate control for performing gate control for entry / exit in the entrance unit 10 and the entry unit 30. A unit 80 and a main control unit 90 that controls each unit are provided.

Further, the database 40 is to calculate information about the characteristics of the parked vehicle, that is, the vehicle data storage unit 41 for storing the feature amount extracted about the vehicle, and the extraction data of the characteristics of the vehicle as needed. It has a feature amount calculation data storage unit 42 for storing various data to be captured by.

Further, the feature amount extraction unit 60 reads the vehicle number, that is, the vehicle number reading unit 61 that reads the numbers and the like displayed on the license plate, and the feature amount other than the vehicle number among the feature amounts. It has a feature amount calculation unit 62 for calculating. In the present embodiment, the feature amount of the vehicle extracted by the image extraction by the feature amount calculation unit 62 or the calculation of the image information is referred to as the feature amount related to the vehicle image.

Here, in particular, the parking lot management device 100 functions as an entry / exit management unit that manages entry / exit of the parking lot as a part of general management of the parking lot, and at the same time, performs characteristic parts of each vehicle in the processing at the time of entry / exit of the vehicle. It functions as a vehicle recognition device for identifying and enabling vehicle recognition. Specifically, among the parking lot management devices 100, the vehicle recognition device VD is provided by the image data acquisition unit 50, the feature amount extraction unit 60, the vehicle identification unit 70, and the feature amount calculation data storage unit 42 of the database 40. By appropriately operating each unit under the control of the main control unit 90, the unique feature amount of each vehicle is extracted and each vehicle can be identified. In particular, in the present embodiment, the feature amount extraction unit 60 calculates not only the vehicle number reading unit 61 that reads the vehicle number, which is the most representative feature amount, but also the feature amount related to the vehicle image, which is a feature amount other than the vehicle number. By providing the feature amount calculation unit 62, for example, even if some vehicle numbers cannot be read by the vehicle number reading unit 61, each vehicle can be identified by the feature amount of the vehicle, that is, the vehicle. It is possible to reliably identify the.

Hereinafter, each part constituting the vehicle recognition device VD described above will be described in more detail. First, in the vehicle recognition device VD, the image data acquisition unit 50 is a place where image data of the vehicle, which is the target of information analysis for vehicle recognition, is received from an external organization, and is, for example, an interface of the parking lot management device 100. It is assumed that it is a data receiving unit or a relay unit that is connected to the data receiving unit and receives image data. The image data acquisition unit 50 appropriately transmits data or the like so that the received data can be processed by the feature amount extraction unit 60. In the present embodiment, the image data acquisition unit 50 has an image pickup function in the entrance unit 10 and the entry unit 30, which are external organizations for the vehicle recognition device VD, so that the image data acquisition unit 50 can enter from the entrance unit 10 and the entry unit 30. And obtain the image data of the vehicle at the time of participation.

Of the vehicle recognition device VD, the feature amount extraction unit 60 is configured by reading and executing various programs stored in the feature amount calculation data storage unit 42 in the database 40, for example, in a GPU or the like, and is an image data acquisition unit. By performing necessary image processing on the license plate of the vehicle acquired from the entrance section 10 and the entry section 30 via the 50, information on the characters, numbers, etc. written on the license plate can be obtained. It is an image analysis unit that performs data conversion processing. The image analysis processing in the feature amount extraction unit 60 will be described in detail later, but various known image processing methods are used for specific image analysis methods for extracting various feature amounts for characters, figures, colors, and the like. The method can be applied.

Of the vehicle recognition device VD, the vehicle identification unit 70 is configured by reading and executing various programs stored in the feature amount calculation data storage unit 42 in the database 40, for example, in a CPU or the like, and is configured by the feature amount extraction unit 60. Based on various feature amount data extracted or calculated in, it is determined whether or not sufficient information is obtained to enable the distinction between one vehicle and another vehicle, thereby identifying the vehicle. conduct.

In the vehicle recognition device VD, the feature amount calculation data storage unit 42 is used to form the feature amount extraction unit 60, for example, data for performing various pattern matching for recognizing a vehicle number, a front portion of the vehicle, and the like. As described above, data for matching the shape features of each member existing around the license plate are stored and stored in a database. In addition, various data for capturing the degree of straightness of the straight line portion and the degree of curve of the curved line portion are also stored. That is, these data are appropriately read out in the shape extraction or the like in the feature amount extraction unit 60. In addition, the feature amount calculation data storage unit 42 further determines various threshold values necessary for specifying the vehicle, in addition to the data for pattern matching as described above, in order to configure the vehicle identification unit 70. Data and the like are also stored, and these data are used in the vehicle identification or determination in the vehicle identification unit 70. These data are stored in advance in the database 40. However, by providing a learning function, these data may also be updated.

By providing each of the above parts as a vehicle recognition device VD, the parking lot management device 100 can recognize a vehicle.

Hereinafter, other parts constituting the parking lot management device 100 will be described.

Of the parking lot management device 100, the database 40 is composed of various storage devices and the like for holding various data, and as described above, depending on the purpose of the vehicle data storage unit 41, the feature amount calculation data storage unit 42, and the like. The data is organized and stored. For example, as described above, the vehicle data storage unit 41 stores information regarding the characteristics of the parked vehicle. Specifically, as shown in FIG. 4, feature amount data indicating features such as vehicle numbers, license plate shades, and various shapes extracted from the vehicle are stored, and these data are stored when the vehicle participates. It is organized and stored so that it can be read out as appropriate when checking the vehicle at. In addition, various data about the vehicle are newly created every time the entrance of the vehicle is confirmed by the entrance section 10. On the other hand, when the participation of the vehicle is confirmed, some data will be deleted. That is, the data in the vehicle data storage unit 41 is updated each time.

The gate control unit 80 sends a command signal for opening / closing the gate for entry / exit at the entrance unit 10 and the entrance unit 30 according to the control of the parking lot management device 100, that is, according to the instruction of the main control unit 90. Create and send to the entrance section 10 and the entry section 30.

The main control unit 90 comprehensively controls the operation of each of the above-mentioned units and is responsible for overall management of the parking lot. In particular, the main control unit 90, as the main part of the parking lot management device 100, operates each part other than the above-mentioned component part of the vehicle recognition device VD in addition to the operation management of the vehicle recognition device VD to operate the vehicle. It manages general matters in parking management, such as entering and exiting the parking lot and grasping the number of vehicles. For example, as a vehicle entrance management unit, the gate control unit 80 is instructed to close the gate to stop until the necessary information about the vehicle to be entered is acquired, and after the information is acquired, the gate is opened. The information of the entered vehicle is stored in the vehicle data storage unit 41 of the database 40 while prompting the entry of the vehicle. Further, as the vehicle participation management unit, the information of the vehicle to be entered is collated with the information in the vehicle data storage unit 41 of the database 40, and the participation is permitted. That is, it functions as a collation unit for collating vehicle data. Also during this period, the gate control unit 80 is made to perform the opening / closing operation. However, among the operations during this period, the vehicle recognition device VD is responsible for acquiring information indicating the characteristics of the vehicle to enter and the vehicle to enter.

Hereinafter, a configuration example of the entrance section 10 and the entry section 30 connected to the parking lot management device 100 in the parking lot management system 500 will be described.

Of the parking lot management system 500, the entrance section 10 includes a ticket issuing machine 11, a gate device 12, and a first image pickup device 13. The ticket issuing machine 11 is a parking ticket issuing machine that issues a parking ticket from the discharge port 11a. The gate device 12 is, for example, a device that raises and lowers the gate bar 12a according to a command signal from the parking lot management device 100 to indicate whether or not to allow the vehicle to enter. The gate device 12 receives a command signal from the gate control unit 80 of the parking lot management device 100 via the ticket issuing machine 11, and also receives a command signal from the ticket issuing machine 11. The first image pickup device 13 is an image pickup device having a camera unit composed of CMOS or the like, and is arranged in front of the gate bar 12a as shown at least on the front side of the vehicle to be entered. The license plate part and the surrounding part in front of the vehicle are installed in a position where the image can be taken.

Of the parking lot management system 500, the entry unit 30 includes a settlement device 31, a gate device 32, and a second image pickup device 33. The settlement device 31 inserts and discharges the parking ticket from the reception port 31a and performs various processes necessary for settlement of the parking fee. The gate device 32 is, for example, a device that raises and lowers the gate bar 32a according to a command signal from the gate control unit 80 of the parking lot management device 100 to indicate whether or not to allow the vehicle to enter. The gate device 32 receives a command signal from the parking lot management device 100 via the checkout device 31, and also receives a command signal from the checkout device 31. The second image pickup device 33 is an image pickup device having a camera unit composed of CMOS or the like, and is arranged in front of the gate bar 32a as shown in the figure, so that at least the front side of the vehicle to participate in is at least. The license plate part and the surrounding part in front of the vehicle are installed in a position where the image can be taken.

The image data acquired by the first image pickup device 13 of the entrance unit 10 and the second image pickup device 33 of the entry unit 30 are transmitted to the parking lot management device 100, and the transmitted image data is acquired by the image data acquisition unit 50. This enables image data analysis in the vehicle recognition device VD.

Hereinafter, an example of the processing content in the vehicle recognition device VD according to the present embodiment will be described in more detail with reference to FIG.

Here, as described above, it is assumed that the feature amount extraction unit 60 of the vehicle recognition device VD is composed of a vehicle number reading unit 61 and a feature amount calculation unit 62. That is, the vehicle number reading unit 61 reads the vehicle number of the vehicle to be entered and exited when the vehicle enters and exits, and the feature amount calculation unit 62 calculates the feature amount related to the vehicle image which is a feature amount other than the vehicle number. .. Further, as an example here, as shown in the figure, the feature amount calculation unit 62 includes a plate shade identification unit 62a that identifies the shade of the license plate, and a plate shape identification unit 62b that identifies the shape of the license plate and the contour. It is assumed that it is composed of a peripheral shape identification unit 62c that identifies the peripheral shape of the license plate.

As an example, as shown in FIG. 3A, the vehicle number reading unit 61 has a vehicle number, that is, a land transportation branch office, from the image data GI about the license plate NP among the image data acquired by the first image pickup device 13 and the like. (Or automobile inspection registration office) The vehicle number is divided into four partial images P1 to P4 of the place name, classification number, kana character and serial designation number, and the character and number information is read from each partial image P1 to P4. Performs the process of converting to data as information.

As an example, the plate shade identification unit 62a and the plate shape identification unit 62b have a color shade pattern and a characteristic shape from the image data GI of the license plate NP having a pattern as illustrated in FIG. 3 (B). Data is extracted for. For example, when a pink floral pattern is applied on a white license plate, a floral pattern is extracted based on the difference in shade, and the extracted information is information for identifying the vehicle. Treated as. Alternatively, if there is a mark on the license plate that has five circles, this shape is extracted, and the extracted information is treated as information for identifying the vehicle. If the contour shape of the license plate is characteristic, for example, even if a part of the license plate is deformed, this shape is extracted and treated as information for identifying the vehicle.

As an example, the peripheral shape identification unit 62c extracts data from the image data GP for a portion other than the license plate for a portion having a characteristic shape, as illustrated in FIG. 3C. For example, the shape features of the entire surface around the license plate and each partial image of the bumper, bonnet, headlight, etc. are extracted, and the characteristic shape peculiar to the vehicle is selected from these. The selected information is treated as information for identifying the vehicle.

FIG. 4A is a diagram showing an example of a list of vehicle data read by extraction by the feature amount extraction unit 60, and in particular, FIG. 4A is a diagram shown by the vehicle number reading unit 61. It mainly shows the vehicle number information, which is the information about the read vehicle number. Further, FIG. 4B is a feature amount as information indicating a feature amount related to a vehicle image, which is information outside the vehicle number extracted or calculated by the feature amount calculation unit 62 among the vehicle data shown in FIG. 4A. It is a figure which showed an example of data. That is, FIG. 4B shows an example of a list of data such as shapes and shading patterns read by the calculation by the feature amount calculation unit 62.

Here, in recent years, the number of license plates having new characters and patterns, such as so-called local numbers, is increasing. That is, the design of license plates is diversifying. For example, in the case of a conventional license plate without a pattern or the like, in reading by the vehicle number reading unit 61 as illustrated in FIG. 3A, each number or character can be discriminated with a relatively high probability, and the recognition rate is high. It is considered to be. Further, it is considered that the matching rate of the matching pattern at that time is also high, that is, the possibility of misreading is low. On the other hand, in the case of a license plate containing a pattern as illustrated in FIG. 3B, information other than the normal number information may become noise, and the reading rate of the vehicle number reading unit 61 may decrease. That is, it is considered that the possibility of unread or misreading the car number increases. That is, as illustrated by "? (Question mark)" in FIG. 4 (A), depending on the vehicle, a part of the number information may be missing due to unread vehicle number. Further, in such a case, it is considered that some of the parts where the vehicle number can be read may be misread.

Therefore, in the present embodiment, as described above, the feature amount extraction unit 60 is required to specify the vehicle in the vehicle number reading unit 61 by providing the feature amount calculation unit 62 in addition to the vehicle number reading unit 61. Even if sufficient information cannot be obtained, as additional information about the vehicle to supplement this, information indicating the feature amount related to the above-mentioned vehicle image can be acquired.

Specifically, as shown as the feature amount data in FIG. 4B, the feature amount calculation unit 62 calculates the vehicle whose feature amount information is "Yes" in the vehicle data of FIG. 4A. Regarding the information acquired by performing the data, the position where the data was acquired, the type of the acquired feature, the content of the feature amount data according to the type, and the like are listed.

It is conceivable that the extent to which the feature amount data is acquired is appropriately specified in advance by the procedure, the threshold value for the recognition rate, and the like. For example, as shown in the list in FIG. 4A, for the second vehicle in the vehicle data, all the vehicle number information read by the vehicle number reading unit 61 is acquired. In this case, it is highly possible that the accuracy is high, and in such a case, it is considered that it is not necessary to acquire the information using the feature amount calculation unit 62. For such a vehicle, the image processing may be terminated without acquiring the data by the feature amount calculation unit 62. By doing so, quick entry / exit processing can be performed.

Hereinafter, an example of a series of processes of the parking lot management device 100 at the time of entering the parking lot of the vehicle will be described with reference to the flowchart of FIG.

First, when the vehicle is detected and imaged in the entrance section 10, the captured image data is transmitted to the image data acquisition section 50 of the vehicle recognition device VD to acquire image data about the license plate and its surroundings. (Step S101). In step S101, the vehicle recognition device VD that has acquired the image data first reads the vehicle number in the vehicle number reading unit 61 of the feature amount extraction unit 60 (step S102), and the vehicle number reading is sufficient in step S102. The vehicle identification unit 70 confirms whether or not the number information sufficient for vehicle identification has been obtained (step S103). Here, the determination criteria, that is, the threshold values in step S103 can be set in various ways, but for example, it is assumed that the vehicle number reading is sufficient when data can be acquired for all four partial images P1 to P4 in FIG. 3 (A). Can be considered. When it is determined in step S103 that the vehicle number has been read by the vehicle identification unit 70 (step S103: Yes), the vehicle identification unit 70 registers the feature amount of the specified vehicle in the vehicle data storage unit 41 of the database 40. Therefore, a transmission to the effect that registration is possible to the main control unit 90 is performed. The main control unit 90 stores the information acquired by the vehicle recognition device VD in a predetermined area of the vehicle data storage unit 41 of the database 40 (step S104), and serves as an entry / exit management unit to allow the vehicle to enter the gate. The gate control unit 80 is made to perform the opening operation (step S108).

On the other hand, in step S103, for example, when it is determined that data has not been acquired in any of the four partial images P1 to P4 in FIG. 3A and the vehicle number has not been read (step S103: No). The feature amount calculation unit 62 of the feature amount extraction unit 60 performs the feature amount calculation process (step S105). That is, the feature amount calculation unit 62 extracts and calculates various feature amounts. Here, regarding the extraction / calculation method of a plurality of types of feature quantities, it is conceivable to appropriately determine the priority order. That is, it is conceivable to select in advance a method for extracting and calculating a feature amount that is likely to be able to identify a unique vehicle, and prioritize these processes. In step S105, for example, when some priority predetermined feature quantities are extracted and calculated, whether or not the vehicle identification unit 70 can sufficiently acquire vehicle-specific information from the acquired feature quantity information. Is determined. That is, the vehicle identification unit 70 appropriately reads information on various threshold values necessary for specifying the vehicle stored in the feature amount calculation data storage unit 42, and the feature amount information acquired in step S105 reaches the threshold value. It is determined whether or not it is (step S106). If it is determined in step S106 that the threshold value has not been reached (step S106: No), the feature amount calculation unit 62 acquires the feature amount data by a new feature amount extraction / calculation method (step S105). In step S106, the above operation is repeated until it is determined that the threshold value has been reached. When it is determined in step S106 that the threshold value has been reached (step S106: Yes), the vehicle identification unit 70 has a main control unit in order to register the feature amount of the specified vehicle in the vehicle data storage unit 41 of the database 40. Send to the effect that registration is possible in 90. The main control unit 90 stores all the information acquired for vehicle identification including the partially missing number information acquired so far in a predetermined area of the vehicle data storage unit 41 of the database 40 (step S107). At the same time, as the entry / exit management unit, the gate control unit 80 is made to perform an operation of opening the gate to allow the vehicle to enter (step S108).

In the above series of processes, as a typical example when it is determined in step S106 that the threshold value has not been reached and the process in step S105 is repeated again, the range around the license plate identified by the peripheral shape identification unit 62c is typical. Can be considered as a method of gradually expanding. For example, in the first step S105, identification is performed only within the license plate, and if the threshold value is not reached, in the second step S105, identification is also performed on the immediate outside of the license plate. If the threshold is still not reached, a method of further expanding the identification range outside the license plate and repeating this operation until the threshold is reached can be considered.

Further, in the above case, when the vehicle identification unit 70 cannot identify the vehicle only by the vehicle number information that can be read by the vehicle number reading unit 61 in step S102, the vehicle identification unit 70 calculates the feature amount in step S105. The vehicle is recognized by using the information of the feature amount related to the vehicle image calculated in the unit 62 to identify the vehicle.

As described above, the vehicle recognition device VD can maintain a state in which each vehicle that has entered can be identified.

Hereinafter, an example of a series of processes of the parking lot management device 100 at the time of entering the parking lot of the vehicle will be described with reference to the flowchart of FIG.

First, when the vehicle is detected and imaged in the participation unit 30, the captured image data is transmitted to the image data acquisition unit 50 of the vehicle recognition device VD to acquire image data about the license plate and its surroundings. (Step S201). In step S201, the vehicle recognition device VD that has acquired the image data first reads the vehicle number in the vehicle number reading unit 61 of the feature amount extraction unit 60 (step S202). In step S202, when the number information, which is the information about the vehicle number, is read, the main control unit 90, as the collation unit, has information matching the read number information in the data of the vehicle data storage unit 41. Vehicle number data collation is performed to confirm whether or not (step S203). If there is a match as a result of the collation in step S203 (step S204: Yes), the main control unit 90 deletes the data in the vehicle data storage unit 41 (step S205), and the entry / exit management unit. As a result, the gate control unit 80 is made to perform an operation of opening the gate in order to allow the vehicle to participate (step S210).

On the other hand, when it is determined that there is no match as a result of the collation in step S203 (step S204: No), the feature amount calculation unit 62 of the feature amount extraction unit 60 performs the feature amount calculation process (step S204: No). Step S206). When the feature amount data is calculated in step S206, the main control unit 90, as a collation unit, confirms whether or not information matching the calculated feature amount data exists in the data of the vehicle data storage unit 41. The feature amount data collation is performed as much as possible (step S207). Specifically, among the vehicle recognition devices VD, the characteristic amount of the vehicle on the vehicle data storage unit 41 of the registered database 40 in the vehicle identification unit 70 and the specified vehicle, that is, the characteristics of the vehicle to participate in. By matching with the quantity, it is determined whether or not it is the same as one vehicle on the vehicle data storage unit 41. If there is a match as a result of the collation in step S207 (step S208: Yes), the vehicle identification unit 70 causes the main control unit 90 to delete the matched information from the vehicle data storage unit 41 of the database 40. Send to that effect. The main control unit 90 deletes the data in the vehicle data storage unit 41 (step S209), and causes the gate control unit 80 to open the gate to allow the vehicle to enter as an entry / exit management unit. (Step S210). On the other hand, when it is determined that there is no match as a result of the collation in step S207 (step S208: No), the vehicle recognition device VD determines that the vehicle cannot be identified and cannot be determined, and the passenger of the vehicle or Various communication operations and notification operations are performed in order to notify the parking lot manager and the like that the participation cannot be permitted because the determination cannot be made (step S211).

Hereinafter, a modified example of the above embodiment will be described with reference to FIG. 7 and the like. First, in the above, the vehicle recognition device VD is incorporated as a partial configuration of the parking lot management device 100 by using each part constituting each part of the parking lot management device 100. For example, FIG. 7 shows. As described above, in the parking lot management device 100, the vehicle recognition device VD may exist independently of other parts. Specifically, in the illustrated example, the parking lot management device 100 is composed of a vehicle recognition device VD that performs vehicle recognition and a parking management unit PM that performs general parking management other than vehicle recognition. In addition to the above-described configuration shown in FIG. 1 and the like, the vehicle recognition device VD further includes a vehicle recognition control circuit 92 that is composed of a CPU or the like and controls the operation of each part in an integrated manner. In other words, the vehicle recognition control circuit 92 controls the vehicle recognition among the various controls carried out by the main control unit 90 in the case of FIG. 1. On the other hand, the parking management unit PM has each unit other than the vehicle recognition device VD for parking lot management, and further has a data reception unit 20 and a parking management main control unit 91. The data reception unit 20 functions as an interface unit that communicates with external organizations such as the entrance unit 10 and the entry unit 30, and is connected to the image data acquisition unit 50 of the vehicle recognition device VD to receive an image of the vehicle received from the external organization. The data is transmitted to the image data acquisition unit 50. The parking management main control unit 91 controls matters other than vehicle recognition among the various controls carried out by the main control unit 90 in the case of FIG. 1. Even with the above configuration, the same operation as in the above-mentioned case can be established. In this case, the vehicle recognition device VD performs image processing exclusively for vehicle recognition, while the parking management unit PM does not perform vehicle recognition processing, and the entrance is based on the analysis result of the vehicle recognition device VD. It is a mode to manage the data of the vehicle that has been used, collate the data of the vehicle that is going to participate, and perform general parking management such as permission to enter and exit. That is, the parking management unit PM functions as an entry / exit management unit that manages entry / exit of the parking lot based on the feature amount of the vehicle acquired by the vehicle recognition device VD.

Further, in the case of the above configuration, as a further modification, as shown in FIG. 8, a vehicle recognition device VD is provided on the side of the first image pickup device 13 and the second image pickup device 33 to recognize the vehicle. Data acquisition may be performed by the entrance section 10 or the entry section 30. Further, as shown in FIG. 9, it is conceivable that the ticket issuing machine 11 and the payment device 31 are provided with the vehicle recognition device VD.

As described above, in the present embodiment, when the vehicle is specified, not only the vehicle number but also the feature amount peculiar to the vehicle is used, so that, for example, some vehicle numbers cannot be read. However, it is possible to reliably identify the vehicle by the feature amount of the vehicle. That is, even when a new character, a pattern number, or the like is added, it is possible to immediately respond to this by using the feature amount of the vehicle, and it is possible to improve the recognition rate. As a result, for example, in parking lot management, it is possible to more reliably determine whether or not to allow a vehicle to pass through when entering or exiting the parking lot.

[Second Embodiment]
Hereinafter, the parking lot management system according to the second embodiment will be described with reference to FIG. 10 and the like. The parking lot management system 600 of the present embodiment is a modified example of the parking lot management system 500 (see FIG. 1) of the first embodiment, and the parts having the same reference numerals or the names, which are not particularly described, are not particularly described. , The same as in the case of the first embodiment.

In the parking lot management system 600 of the present embodiment, the parking lot to be managed is, for example, a parking lot for residents of an apartment house or a parking lot for a vehicle having a regular contract such as monthly. It differs from the case of the first embodiment in that it is a system for managing the entry / exit of a pre-registered vehicle. Therefore, in the present embodiment, in the database unit 240, the registered vehicle data storage unit 241 is stored instead of the vehicle data storage unit 41 in FIG. 1. That is, the vehicle data storage unit 41 of FIG. 1 is updated every time a vehicle enters and exits, but the registered vehicle data storage unit 241 of the present embodiment has a number such that the contractor of the parking lot or the vehicle changes. The difference is that the data is changed only when there is a possibility of changing the plate, and in principle, it is not changed. Further, the entrance section 210 and the entry section 230 are different in that they do not have the ticket issuing machine 11, the checkout device 31, and the second image pickup device 33 in FIG. 1.

Here, in the above case, the data of each vehicle in the registered vehicle data storage unit 241 may include the characteristics of the license plate of the contract vehicle in advance. For example, in the case of a license plate containing a design, the characteristics that are likely to occur from that design may be stored as unique data, and conversely, the parts of the vehicle number that are likely to be misread are not registered from the beginning. In this way, it may be excluded from the judgment target at the time of collation. By selecting the determination information according to each vehicle, it is possible to identify the correct vehicle with a very high probability.

Hereinafter, an example of a series of processes of the parking lot management device 200 at the time of entering the vehicle from the parking lot will be described with reference to the flowchart of FIG.

First, when the vehicle is detected and imaged in the entrance unit 210, the imaged image data is transmitted to the image data acquisition unit 50 of the vehicle recognition device VD to acquire image data about the license plate and its surroundings. (Step S301). In step S301, the vehicle recognition device VD that has acquired the image data performs vehicle number reading and feature amount calculation in the vehicle number reading unit 61 and the feature amount calculation unit 62 of the feature amount extraction unit 60 (step S302). .. In step S302, when the number information which is the information about the vehicle number is read and the feature amount data is calculated, the main control unit 90 serves as the collation unit and the information matching the read number information is the vehicle data. Vehicle number data collation is performed to confirm whether or not the data exists in the data of the storage unit 41 (step S303). If there is a match as a result of the collation in step S303 (step S304: Yes), the main control unit 90 causes the gate control unit 80 to open the gate in order to allow the vehicle to enter (step S304). S307).

On the other hand, when it is determined as a result of the collation in step S303 that there is no match (step S304: No), the main control unit 90 performs a process of requesting another collation means (step S305). As for the processing in step S305, various existing methods can be considered, and for example, a button part (not shown) for pressing the password and a voice part for prompting the user to press the password are prepared at the entrance. It is conceivable to set up a facility to enable authentication with the entrance key.

When the collation operation is performed as a result of the collation in step S305 (step S306: Yes), the main control unit 90 causes the gate control unit 80 to perform an operation of opening the gate in order to allow the vehicle to enter (step S307). .. On the other hand, if the collation operation is not performed as a result of the collation in step S305 (step S306: No), the vehicle cannot be identified and the determination cannot be made, and the determination cannot be made to the passenger of the vehicle, the manager of the parking lot, or the like. Therefore, various communication operations and notification operations are performed in order to notify that participation cannot be permitted (step S308).

Regarding the entry operation, the above processing is not provided. For example, various sensors (not shown) are provided in the entry section 230, and the sensor automatically detects that the vehicle is approaching the entry section 230. It is conceivable to open the gate in a specific manner. Further, the participation unit 230 itself may be omitted.

In addition, in this embodiment, it may be confirmed whether the vehicle identification unit 70 can identify the registered vehicles to the extent that they are not confused with each other. For example, whether or not the information about the feature amount extracted for vehicle identification at the time of registration of a new contract vehicle is sufficient to distinguish it from other vehicles with the already registered vehicle. It is conceivable to confirm.

As described above, also in the present embodiment, when specifying a vehicle, it is possible to reliably specify the vehicle by using not only the vehicle number but also the feature amount peculiar to the vehicle.

〔others〕
The present invention is not limited to each of the above embodiments, and can be carried out in various embodiments without departing from the gist thereof.

First, in the above embodiment, the acquired license plate image data is classified into four (place name, classification number, kana character and serial designation number of the Land Transport Office) and converted into data. Is not limited to this, and various aspects are possible, and only a part of these may be targeted for data conversion.

Further, in the above, the feature amount calculation unit 62 is composed of a plate shading identification unit 62a, a plate shape identification unit 62b, and a peripheral shape identification unit 62c, but other extraction / calculation is performed. For example, various information related to color such as lightness and saturation of color may be extracted and calculated, and the peripheral information of the number plate may be extracted and calculated not only by shape but also by shading and color. May be good. Alternatively, it is conceivable to use only some of these and not the others.

In addition, various aspects can be considered for the method of determining the threshold value and the determination criteria. For example, in the first embodiment described above, it is required that all four items of the place name, classification number, kana character, and serial designation number of the Land Transport Office match, but the present invention is not limited to this, and for example, the serial designation number is included. If up to three items match, it may be treated as matching.

Further, in the second embodiment, the vehicle identification unit 70 confirms whether or not the identification can be performed to the extent that no confusion occurs between the registered vehicles. However, the same determination criteria are also used in the first embodiment. The threshold value for identification may be set with. In other words, when acquiring information on the feature amount to identify a new vehicle to enter, admission is permitted when sufficient information is obtained to distinguish it from the vehicles that have entered so far. It may be a thing.

Further, in the above description, the first embodiment focuses on temporary parking, and the second embodiment focuses on dedicated parking and regular parking. However, in a parking lot that combines these. , The invention of the present application may be applied. In addition, it can be widely applied not only to pay parking lots but also to parking lots that manage based on vehicle numbers.

Further, in the above configuration, the vehicle recognition device VD may have a learning function. That is, as a means of converting the commonality of the acquired various feature quantities into data and further improving the recognition rate, for example, by finding a common item for the symbols of the same local number, it is determined which local number it is. It is conceivable to make it possible and to handle it as one of the vehicle-specific features.

It should be noted that the information on the individual identification of the vehicle based on the acquired various feature quantities can be used in various modes, for example, by associating it with the number of the parking ticket issued by the ticket issuing machine 11 to enable pre-payment. It is possible.

10 ... entrance, 11 ... ticketing machine, 11a ... exit, 12 ... gate device, 12a ... gate bar, 13 ... imaging device, 20 ... data reception, 30 ... entry, 31 ... settlement device, 31a ... reception, 32 ... Gate device, 32a ... Gate bar, 33 ... Imaging device, 40 ... Database, 41 ... Vehicle data storage unit, 42 ... Feature amount calculation data storage unit, 50 ... Image data acquisition unit, 60 ... Feature amount extraction unit, 61 ... Vehicle number reading unit, 62 ... Feature amount calculation unit, 62a ... Plate shade identification unit, 62b ... Plate shape identification unit, 62c ... Peripheral shape identification unit, 70 ... Vehicle identification unit, 80 ... Gate control unit, 90 ... Main control unit , 91 ... Parking management main control unit, 92 ... Vehicle recognition control circuit, 100 ... Parking lot management device, 200 ... Parking lot management device, 210 ... Entrance section, 230 ... Participation section, 240 ... Database section, 241 ... Registered vehicle data Storage unit, 500 ... Parking lot management system, 600 ... Parking lot management system, GI, GP ... Image data, NP ... Number plate, P1-P4 ... Partial image, PM ... Parking management unit, VD ... Vehicle recognition device

Claims (6)

The image data acquisition unit that acquires the image data of the vehicle,
A feature amount extraction unit that extracts a feature amount unique to the vehicle including the vehicle number from the image data acquired by the image data acquisition unit, and a feature amount extraction unit.
A vehicle identification unit that identifies a vehicle based on the information extracted by the feature amount extraction unit is provided .
The feature amount extraction unit is a vehicle recognition device that extracts a pattern on a license plate as a feature amount other than the vehicle number . The vehicle recognition device according to claim 1, wherein the feature amount extraction unit extracts the design based on the difference in the color shade of the license plate . The feature amount extraction unit has a vehicle number reading unit that reads a vehicle number as a feature amount to be extracted.
The vehicle specifying unit uses the information of the symbol to identify the vehicle when the vehicle cannot be specified only by the vehicle number information that can be read by the vehicle number reading unit, according to claims 1 and 2 . The vehicle recognition device according to any one of the items. The vehicle specifying unit determines whether or not it is the same as one vehicle on the database by matching the feature amount of the vehicle on the registered database with the feature amount of the specified vehicle. The vehicle recognition device according to any one of 1 to 3 . The vehicle recognition device according to any one of claims 1 to 4 .
A parking lot management system including an entry / exit management unit that manages entry / exit of a parking lot based on a vehicle feature amount acquired by the vehicle recognition device. The vehicle identification unit registers the feature amount of the specified vehicle in the database when the vehicle enters, and matches the feature amount of the vehicle on the registered database with the feature amount of the specified vehicle when the vehicle enters. By, it is determined whether or not the same vehicle exists, and
The parking lot management system according to claim 5 , wherein the entry / exit management unit permits entry when it is determined in the vehicle identification section that the same vehicle exists in the database when the vehicle enters.

Images