A license plate recognition method based on panoramic parking lot

Abstract

The license plate recognition method based on the panoramic parking lot disclosed in the present invention obtains the fisheye camera 1 to the fisheye camera 2 through the driving trajectory of the vehicle A. n The method of identifying the license plate information and comparing it to obtain the optimal license plate information can avoid misidentification caused by various interferences, thereby solving the problems existing in the prior art. The present invention includes: firstly detecting that vehicle A enters the fisheye camera m The monitoring range identifies the license plate information of vehicle A. m and store it in the data processing center; then use the body color and license plate number of vehicle A as the judgment data to obtain the driving trajectory of vehicle A; then obtain the fisheye camera 1 to the fisheye camera according to the driving trajectory n , call the fisheye camera 1~fisheye camera in the data processing center n Corresponding license plate information A1~license plate information A n ; Finally, license plate information A1 ~ license plate information A n The license plate number in is compared, the real license plate number is obtained according to the comparison rules, and the real license plate number is replaced with the license plate number in the license plate information A.

Description

Technical Field

[0001] The present invention relates to the technical field of parking lots, and in particular to a license plate recognition method based on a panoramic parking lot. Background Art

[0002] A parking lot is a place for vehicles to park. There are simple parking lots with only parking spaces but no management or fees, and there are paid parking lots with entrance and exit gates, parking attendants, and timekeepers.

[0003] In a parking lot, the license plate recognition of vehicles parked or driving in the parking lot is a relatively important technical issue. If the license plate information of the vehicle can be accurately obtained, many services for the vehicle can be realized, such as the guidance service for the vehicle, which can accurately guide the vehicle to the reserved or other suitable parking space, and can also realize the monitoring and protection function services of the vehicle. In the prior art, the license plate of the vehicle is generally only recognized when the vehicle enters and exits the parking lot. Few parking lots can realize the license plate recognition of vehicles in the parking lot. Some traditional parking lots use two technical means of parking space shooting and lane shooting for license plate recognition. Among them, the parking space shooting method is not well adapted to some construction environments, such as side parking spaces, parent-child parking spaces, etc.; and the channel shooting is easily disturbed by the close following of the car, and the perception of the vehicle moving in a small range in the parking space is also poor. Therefore, it is difficult to meet the needs of existing users.

[0004] Therefore, in view of the above situation, how to improve the existing technology so as to solve the above problems has become an important technical problem that needs to be solved urgently by those skilled in the art. Summary of the invention

[0005] The present invention discloses a license plate recognition method based on a panoramic parking lot, which obtains the fisheye camera 1 to the fisheye camera 2 through the driving trajectory of the vehicle A. n The method of identifying the license plate information and comparing it to obtain the optimal license plate information can avoid misidentification caused by various interferences, thereby solving the problems existing in the prior art.

[0006] The license plate recognition method based on a panoramic parking lot provided by the present invention is used in a panoramic parking lot, wherein the panoramic parking lot comprises a plurality of fisheye cameras, a network module and a data processing center; a monitoring area is defined in the panoramic parking lot, and the number of the fisheye cameras is based on the number of cameras that can cover the monitoring area, including:

[0007] From the vehicle entering to leaving the panoramic parking lot, fisheye camera 1 to fisheye camera n Perform license plate recognition on vehicles in sequence, n≥1, 1≤m≤n;

[0008] S1: Vehicle A is detected entering the fisheye cameram The monitoring range identifies the license plate information A of the vehicle A. m and store it in the data processing center;

[0009] S2: using the body color and license plate number of the vehicle A as determination data, obtaining the driving trajectory of the vehicle A;

[0010] S3: Obtaining fisheye camera 1 to fisheye camera 2 according to the driving trajectory n , call the fisheye camera 1~fisheye camera in the data processing center n Corresponding license plate information A1~license plate information A n ;

[0011] S4: License plate information A1 to license plate information A n The license plate number in the license plate information A is compared, and the real license plate number is obtained according to the comparison rule, and the real license plate number is used to replace the license plate number in the license plate information A.

[0012] Preferably,

[0013] The license plate information A m Includes: the license plate number of vehicle A, and at least one of the vehicle color, driving direction or vehicle trajectory.

[0014] Preferably,

[0015] Read vehicle A entering the fisheye camera m The starting point of the monitoring range of the fisheye camera m The identification point for the next photo identification of vehicle A;

[0016] The direction from the starting point to the identification point is the driving direction of vehicle A.

[0017] Preferably,

[0018] The step S1 comprises:

[0019] Read vehicle A entering the fisheye camera m The entry and exit points of the monitoring area;

[0020] The line segment between the entry point and the exit point is evenly spaced with h isolation points, where h≥2;

[0021] Identify the license plate information of vehicle A at the entry point, isolation point and exit point, and obtain h+2 license plate information;

[0022] Read the license plate number from the h+2 license plate information, and take the license plate information corresponding to the license plate number that appears most times as license plate information A.

[0023] Preferably,

[0024] The h is an even number.

[0025] Preferably,

[0026] The determination data based on the body color and license plate number of the vehicle A includes:

[0027] If the fisheye camera m If there is only one vehicle in the monitoring range, the body color of the vehicle A is used as the determination data; otherwise, the body color of the vehicle A and the approximate license plate number are used as the determination data;

[0028] The approximate license plate number is a license plate number with a similarity of not less than 85%.

[0029] Preferably,

[0030] The step S4 comprises:

[0031] License plate information A1~License plate information A n The license plate number in the license plate information A is compared, and the license plate with the most occurrences is taken as the real license plate number, and the real license plate number is used to replace the license plate number in the license plate information A.

[0032] Preferably,

[0033] The license plate that appears the most times as the real license plate number includes:

[0034] If the number of license plates that appear the most times is more than one, an abnormal message will be sent to the data processing center and manual processing will be notified. Otherwise, the license plate that appears the most times will be used as the real license plate number.

[0035] The license plate recognition method based on panoramic parking lot provided by the present invention comprises: firstly detecting that vehicle A enters the fisheye camera m The monitoring range identifies the license plate information of vehicle A. m and store it in the data processing center; then use the body color and license plate number of vehicle A as the judgment data to obtain the driving trajectory of vehicle A; then obtain the fisheye camera 1 to the fisheye camera according to the driving trajectory n , call the fisheye camera 1~fisheye camera in the data processing center n Corresponding license plate information A1~license plate information A n ; Finally, license plate information A1 ~ license plate information A n The real license plate number is obtained according to the comparison rules, and the real license plate number is replaced with the license plate number in the license plate information A. The fisheye camera 1 to the fisheye camera 2 are obtained through the driving trajectory of vehicle A. nThe license plate information is identified and compared to obtain the optimal license plate information. The license plate recognition method based on the panoramic parking lot of the present invention can avoid misrecognition caused by various interferences, thereby solving the problems existing in the prior art. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0037] Figure 1 It is a flow chart of a first embodiment of a license plate recognition method based on a panoramic parking lot of the present invention;

[0038] Figure 2 It is a flow chart of a second embodiment of the license plate recognition method based on a panoramic parking lot of the present invention;

[0039] Figure 3 It is a flow chart of a third embodiment of the license plate recognition method based on a panoramic parking lot of the present invention;

[0040] Figure 4 This is a flow chart of a fourth embodiment of a method for license plate recognition based on a panoramic parking lot of the present invention. DETAILED DESCRIPTION

[0041] The present invention discloses a license plate recognition method based on a panoramic parking lot, which obtains fisheye camera 1 to fisheye camera 2 through the driving trajectory of vehicle A. n The method of identifying the license plate information and comparing it to obtain the optimal license plate information can avoid misidentification caused by various interferences, thereby solving the problems existing in the prior art.

[0042] The following is a clear and detailed description of the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. Please refer to Figure 1 , which is a first embodiment of the license plate recognition method based on a panoramic parking lot provided by the present invention, is used in a panoramic parking lot, wherein the panoramic parking lot includes a plurality of fisheye cameras, a network module, and a data processing center; a monitoring area is defined in the panoramic parking lot, and the number of the fisheye cameras is based on the number of cameras that can cover the monitoring area, including:

[0043] From the vehicle entering to leaving the panoramic parking lot, fisheye camera 1 to fisheye cameran Carry out license plate recognition on vehicles one by one, n≥1, 1≤m≤n;

[0044] S1: Vehicle A is detected entering the fisheye camera m The monitoring range identifies the license plate information A of the vehicle A. m and store it in the data processing center;

[0045] First of all, it should be noted that fisheye cameras have a relatively large shooting detection range. By installing them at a relatively high position, they can cover the range from the lane to the parking space. After vehicle A enters the parking lot from the parking lot entrance, the first fisheye camera will monitor vehicle A. In relatively low-end cases, there may be only one fisheye camera monitoring vehicle A before vehicle A stops in the parking space. In most cases, from the time vehicle A enters the parking lot to the time it parks in the parking space or leaves the parking lot, it will be monitored by multiple fisheye cameras.

[0046] Then when vehicle A enters the fisheye camera m The monitoring range of the fisheye camera m Directly below or away from the fisheye camera m Monitoring range of fisheye camera m Vehicle A will be monitored all the time, and license plate information A will be obtained m And stored in the data processing center.

[0047] The above license plate information A m It includes: the license plate number of vehicle A, and at least one of the vehicle color, driving direction or vehicle trajectory. It must be pointed out that the license plate information A m The license plate number of vehicle A must be included, and the vehicle color, driving direction or vehicle trajectory can be selected according to the specific situation. The driving direction can be determined by the following methods: Read the vehicle A entering the fisheye camera m The starting point of the monitoring range of the fisheye camera m The identification point for the next photo identification of vehicle A; the direction from the starting point to the identification point is the driving direction of vehicle A.

[0048] Obtained license plate information A m It can be stored in the data processing center so that other fisheye cameras can call it.

[0049] S2: using the body color and license plate number of the vehicle A as determination data, obtaining the driving trajectory of the vehicle A;

[0050] Through fisheye camera m, the body color and license plate number of vehicle A can be obtained, and the driving trajectory of vehicle A can be obtained based on the body color and license plate number of vehicle A.

[0051] It should be noted here that the judgment of the body color of vehicle A by different fisheye cameras is not much different and can be considered consistent. However, there are cases where multiple vehicles of the same color appear in the monitoring range of the same fisheye camera at the same time, so it is necessary to combine the license plate number for judgment. Since there will be certain errors in the recognition of the license plate number of vehicle A by different fisheye cameras, the approximate license plate number can be used as the judgment standard. The above-mentioned approximate license plate number can be a license plate number that differs by only one digit. It must be pointed out that since the probability of vehicles of the same color and similar license plate numbers appearing in a fisheye camera at the same time is extremely low, this situation can be treated as non-existent in actual operation and is not limited here.

[0052] S3: Obtaining fisheye camera 1 to fisheye camera 2 according to the driving trajectory n , call the fisheye camera 1~fisheye camera in the data processing center n Corresponding license plate information A1~license plate information A n ;

[0053] S4: License plate information A1 to license plate information A n The license plate number in the license plate information A is compared, and the real license plate number is obtained according to the comparison rule, and the real license plate number is used to replace the license plate number in the license plate information A.

[0054] License plate information A1~License plate information A n The license plate number in is compared, and the real license plate number is obtained according to the comparison rules. It should be noted that the fisheye camera has a high success rate in recognizing the license plate of vehicle A. Therefore, after multiple fisheye cameras recognize vehicle A, the accurate license plate number, that is, the real license plate number, can be easily obtained by comparing the recognition results. Finally, the real license plate number can be replaced with the license plate number in the license plate information A.

[0055] The license plate recognition method based on panoramic parking lot provided by the present invention comprises: firstly detecting that vehicle A enters the fisheye camera m The monitoring range identifies the license plate information of vehicle A. m and store it in the data processing center; then use the body color and license plate number of vehicle A as the judgment data to obtain the driving trajectory of vehicle A; then obtain the fisheye camera 1 to the fisheye camera according to the driving trajectory n , call the fisheye camera 1~fisheye camera in the data processing center n Corresponding license plate information A1~license plate information A n ; Finally, license plate information A1 ~ license plate information A nThe real license plate number is obtained according to the comparison rules, and the real license plate number is replaced with the license plate number in the license plate information A. The fisheye camera 1 to the fisheye camera 2 are obtained through the driving trajectory of vehicle A. n The license plate information is identified and compared to obtain the optimal license plate information. The license plate recognition method based on the panoramic parking lot of the present invention can avoid misrecognition caused by various interferences, thereby solving the problems existing in the prior art.

[0056] The first embodiment of the license plate recognition method based on a panoramic parking lot of the present invention is introduced above. The second embodiment of the license plate recognition method based on a panoramic parking lot of the present invention is introduced below. Figure 2 The second embodiment of the license plate recognition method based on a panoramic parking lot of the present invention refines step S1 in the first embodiment of the license plate recognition method based on a panoramic parking lot of the present invention, and the other steps are the same, so only the specific refinement process of step S1 is introduced below, and the other steps are not repeated:

[0057] A1: Read vehicle A entering the fisheye camera m The entry and exit points of the monitoring area;

[0058] A2: The line segment between the entry point and the exit point is evenly spaced with h isolation points, where h ≥ 2;

[0059] It should be noted that h is preferably an even number. When h is an even number, it is possible to avoid reading the isolation point when vehicle A is located directly below the fisheye camera, thereby improving the accuracy of the fisheye camera in identifying the license plate number of vehicle A.

[0060] A3: Identify the license plate information of vehicle A at the entry point, isolation point and exit point, and obtain h+2 license plate information;

[0061] A4: Read the license plate number from the h+2 license plate information, and take the license plate information corresponding to the license plate number that appears the most times as license plate information A;

[0062] A5: storing the license plate information A in the data processing center;

[0063] A6: Using the body color and license plate number of the vehicle A as determination data, obtaining the driving trajectory of the vehicle A;

[0064] A7: Obtain fisheye camera 1 to fisheye camera 2 according to the driving trajectory n , call the fisheye camera 1~fisheye camera in the data processing center n Corresponding license plate information A1~license plate information A n ;

[0065] A8: License plate information A1 to license plate information An The license plate number in the license plate information A is compared, and the real license plate number is obtained according to the comparison rule, and the real license plate number is used to replace the license plate number in the license plate information A.

[0066] Obtain fisheye camera 1 to fisheye camera 2 through the driving trajectory of vehicle A. n The license plate information is identified and compared to obtain the optimal license plate information. The license plate recognition method based on the panoramic parking lot of the present invention can avoid misrecognition caused by various interferences, thereby solving the problems existing in the prior art.

[0067] The second embodiment of the license plate recognition method based on a panoramic parking lot of the present invention is introduced above. The third embodiment of the license plate recognition method based on a panoramic parking lot of the present invention is introduced below. Figure 3 The third embodiment of the license plate recognition method based on a panoramic parking lot of the present invention refines step S2 in the first embodiment of the license plate recognition method based on a panoramic parking lot of the present invention, and the other steps are the same, so only the specific refinement process of step S2 is introduced below, and the other steps are not repeated:

[0068] B1: Vehicle A is detected entering the fisheye camera m The monitoring range identifies the license plate information A of the vehicle A. m and store it in the data processing center;

[0069] B2: If the fisheye camera m If there is only one vehicle in the monitoring range, the body color of the vehicle A is used as the judgment data; otherwise, the body color of the vehicle A and the approximate license plate number are used as the judgment data, and the approximate license plate number is a license plate number with a similarity of not less than 85%;

[0070] In the embodiment of the present invention, first determine the fisheye camera m Is there only one vehicle in the monitoring range? If so, the movement trajectory of vehicle A can be determined only by color. Otherwise, the body color and the approximate license plate number need to be combined as judgment data.

[0071] In addition, generally speaking, even if two fisheye cameras have deviations in recognizing the license plate number of the same vehicle, in most cases there is only a single-digit deviation. Therefore, the present invention only considers the case where there is only a single-digit deviation. Generally, a license plate has seven characters, and the case where there is a one-digit deviation means that the similarity is not less than 85%, that is, an approximate license plate number is a license plate number with a similarity of not less than 85%.

[0072] B3: Obtain the driving trajectory of the vehicle A;

[0073] B4: Obtain fisheye camera 1 to fisheye camera 2 according to the driving trajectory n, call the fisheye camera 1~fisheye camera in the data processing center n Corresponding license plate information A1~license plate information A n ;

[0074] B5: License plate information A1 to license plate information A n The license plate number in the license plate information A is compared, and the real license plate number is obtained according to the comparison rule, and the real license plate number is used to replace the license plate number in the license plate information A.

[0075] Obtain fisheye camera 1 to fisheye camera 2 through the driving trajectory of vehicle A. n The license plate information is identified and compared to obtain the optimal license plate information. The license plate recognition method based on the panoramic parking lot of the present invention can avoid misrecognition caused by various interferences, thereby solving the problems existing in the prior art.

[0076] The third embodiment of the license plate recognition method based on a panoramic parking lot of the present invention is introduced above. The fourth embodiment of the license plate recognition method based on a panoramic parking lot of the present invention is introduced below. Figure 4 The fourth embodiment of the license plate recognition method based on a panoramic parking lot of the present invention refines step S4 in the first embodiment of the license plate recognition method based on a panoramic parking lot of the present invention, and the other steps are the same, so only the specific refinement process of step S4 is introduced below, and the other steps are not repeated:

[0077] C1: Vehicle A is detected entering the fisheye camera m The monitoring range identifies the license plate information A of the vehicle A. m and store it in the data processing center;

[0078] C2: using the body color and license plate number of the vehicle A as determination data, obtaining the driving trajectory of the vehicle A;

[0079] C3: Obtain fisheye camera 1 to fisheye camera 2 according to the driving trajectory n , call the fisheye camera 1~fisheye camera in the data processing center n Corresponding license plate information A1~license plate information A n ;

[0080] C4: License plate information A1 to license plate information A n The license plate number in the license plate information A is compared, and the license plate with the most occurrences is taken as the real license plate number, and the real license plate number is used to replace the license plate number in the license plate information A.

[0081] The method of taking the license plate that appears most times as the real license plate number includes: if the number of the license plate that appears most times is more than one, sending abnormal information to the data processing center and notifying manual processing, otherwise taking the license plate that appears most times as the real license plate number.

[0082] It should be noted that if the number of license plates that appear the most times is more than one, it proves that multiple fisheye cameras have made relatively large recognition errors. At this time, it is no longer possible to correct it by itself, and it is necessary to send abnormal information to the data processing center and notify manual processing.

[0083] The license plate recognition method based on panoramic parking lot provided by the present invention comprises: firstly detecting that vehicle A enters the fisheye camera m The monitoring range identifies the license plate information of vehicle A. m and store it in the data processing center; then use the body color and license plate number of vehicle A as the judgment data to obtain the driving trajectory of vehicle A; then obtain the fisheye camera 1 to the fisheye camera according to the driving trajectory n , call the fisheye camera 1~fisheye camera in the data processing center n Corresponding license plate information A1~license plate information A n ; Finally, license plate information A1 ~ license plate information A n The real license plate number is obtained according to the comparison rules, and the real license plate number is replaced with the license plate number in the license plate information A. The fisheye camera 1 to the fisheye camera 2 are obtained through the driving trajectory of vehicle A. n The license plate information is identified and compared to obtain the optimal license plate information. The license plate recognition method based on the panoramic parking lot of the present invention can avoid misrecognition caused by various interferences, thereby solving the problems existing in the prior art.

[0084] A person skilled in the art can understand that all or part of the steps in the above-mentioned embodiment method can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium. The above-mentioned storage medium can be a read-only memory, a disk or an optical disk, etc.

[0085] The above is a detailed introduction to the license plate recognition method based on a panoramic parking lot provided by the present invention. For a person skilled in the art, according to the idea of ​​the embodiments of the present invention, there may be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A license plate recognition method based on a panoramic parking lot, which is used in a panoramic parking lot. The panoramic parking lot includes multiple fisheye cameras, a network module and a data processing center; A monitoring area is defined in the panoramic parking lot, and the number of the fisheye cameras is based on the number of cameras that can cover the monitoring area, characterized in that: include: From the vehicle entering to leaving the panoramic parking lot, fisheye camera 1 to fisheye camera n Carry out license plate recognition on vehicles one by one, n≥1, 1≤m≤n; S1: Vehicle A is detected entering the fisheye camera m The monitoring range identifies the license plate information A of the vehicle A. m and store it in the data processing center; S2: using the body color and license plate number of the vehicle A as determination data, obtaining the driving trajectory of the vehicle A; S3: Obtaining fisheye camera 1 to fisheye camera 2 according to the driving trajectory n , call the fisheye camera 1~fisheye camera in the data processing center n Corresponding license plate information A1~license plate information A n ; S4: License plate information A1 to license plate information A n The license plate number in the license plate information A is compared, and the real license plate number is obtained according to the comparison rule, and the real license plate number is used to replace the license plate number in the license plate information A; The license plate information A m Including: the license plate number of vehicle A, and at least one of the vehicle color, driving direction or vehicle track; The step S1 comprises: Read vehicle A entering the fisheye camera m The entry and exit points of the monitoring area; The line segment between the entry point and the exit point is evenly spaced with h isolation points, where h≥2; Identify the license plate information of vehicle A at the entry point, isolation point and exit point, and obtain h+2 license plate information; Read the license plate number from the h+2 license plate information, and take the license plate information corresponding to the license plate number that appears most times as license plate information A.

2. The license plate recognition method based on panoramic parking lot according to claim 1 is characterized in that: Read vehicle A entering the fisheye camera m The starting point of the monitoring range and the fisheye camera m The identification point for the next photo identification of vehicle A; The direction from the starting point to the identification point is the driving direction of vehicle A.

3. The license plate recognition method based on panoramic parking lot according to claim 1 is characterized in that: The h is an even number.

4. The license plate recognition method based on a panoramic parking lot according to any one of claims 1 to 3, characterized in that: The determination data based on the body color and license plate number of the vehicle A includes: If the fisheye camera m If there is only one vehicle in the monitoring range, the body color of the vehicle A is used as the determination data; otherwise, the body color of the vehicle A and the approximate license plate number are used as the determination data; The approximate license plate number is a license plate number with a similarity of not less than 85%.

5. The license plate recognition method based on panoramic parking lot according to claim 4 is characterized in that: The step S4 comprises: License plate information A1~License plate information A n The license plate number in the license plate information A is compared, and the license plate with the most occurrences is taken as the real license plate number, and the real license plate number is used to replace the license plate number in the license plate information A.

6. The license plate recognition method based on panoramic parking lot according to claim 5 is characterized in that: The license plate that appears the most times as the real license plate number includes: If the number of license plates that appear the most times is more than one, an abnormal message will be sent to the data processing center and manual processing will be notified. Otherwise, the license plate that appears the most times will be used as the real license plate number.

Images