Method and system for improving license plate recognition rate of raw fuel truck

Abstract

This invention discloses a method and system for improving the license plate recognition rate of raw material and fuel trucks, belonging to the field of intelligent transportation technology. The system includes: a checkpoint unit and a vehicle management module; each checkpoint unit includes one main capture camera and zero to multiple auxiliary capture cameras, each capture camera including one main recognition algorithm and zero to multiple auxiliary recognition algorithms; the vehicle management module generates a dynamic license plate number database for each type of checkpoint unit; the checkpoint unit acquires the license plate number recognized by the capture cameras and compares the license plate number with the dynamic license plate number database corresponding to the checkpoint unit to complete the recognition process; the vehicle management module also performs self-learning on the recognition results for use in the next recognition process. It employs multiple capture devices, multiple recognition algorithms, and fuzzy judgment by the vehicle management module, maximizing the recognition rate and reducing manual intervention.

Description

Technical Field

[0001] This invention belongs to the field of intelligent transportation technology, and more specifically, relates to a method and system for improving the license plate recognition rate of raw material and fuel trucks in steel enterprises. Background Technology

[0002] Steel companies typically transport bulk raw materials to raw material yards by train or large truck. When using trucks, license plate recognition is required before the gate can be operated and vehicle logistics and status information can be managed.

[0003] Taking a steel plant's bulk raw materials as an example, approximately 500 to 1000 trucks transport these goods daily. Each truck passes through an average of 5 to 10 gates on its round trip, equivalent to about 5000 identification operations. Due to the dirtiness of the large truck license plates and the unevenness of some gate surfaces, the accuracy of identification is affected. Currently, the vehicle identification efficiency is generally around 95%, resulting in about 250 identification errors per day. This severely impacts vehicle traffic efficiency and causes road congestion. Furthermore, the originally planned unattended gates require significant manual intervention, failing to achieve the intended purpose. Summary of the Invention

[0004] In response to the above-mentioned defects or improvement needs of existing technologies, and in order to infinitely increase the license plate recognition rate of raw material and fuel trucks in steel enterprises to 100%, this invention proposes a method and system for improving the license plate recognition rate of raw material and fuel trucks. It adopts multiple means such as multiple capture devices, multiple recognition algorithms, and fuzzy judgment of vehicle management modules to maximize the recognition rate and reduce manual intervention.

[0005] To achieve the above objectives, according to one aspect of the present invention, a system for improving the license plate recognition rate of raw material trucks is provided, comprising: a checkpoint unit and a vehicle management module;

[0006] Each checkpoint unit includes one main capture camera and zero to more auxiliary capture cameras. Each capture camera includes one main recognition algorithm and zero to more auxiliary recognition algorithms.

[0007] The vehicle management module is used to generate a dynamic license plate number database for each type of checkpoint unit;

[0008] The checkpoint unit is used to acquire the license plate number recognized by the camera and compare the license plate number with the dynamic license plate number database corresponding to the checkpoint unit to complete the recognition process;

[0009] The vehicle management module is also used to learn from the recognition results for use in the next recognition process.

[0010] In some optional implementation schemes, the main capture camera and the auxiliary capture camera are selected from products from different manufacturers, or different models from the same manufacturer. The installation angles of the main capture camera and the auxiliary capture camera are different to improve the recognition range. The main road recognition algorithm of the capture camera uses the manufacturer's own algorithm, while the auxiliary road recognition algorithm uses a custom recognition algorithm based on deep neural networks to reduce the miss rate.

[0011] In some optional implementations, the checkpoint unit includes: a parking lot checkpoint, an entrance checkpoint, a sampling checkpoint, a weighing checkpoint, and a raw material yard checkpoint;

[0012] Wherein, the parking lot entrance checkpoint refers to the checkpoint that all raw material and fuel trucks must pass through; the factory entrance checkpoint refers to the checkpoint that trucks in the parking lot will pass through; the sampling checkpoint refers to the checkpoint that trucks entering the factory entrance checkpoint and having a sampling plan will pass through; the weighing checkpoint refers to the checkpoint that trucks with a weighing plan and having completed the sampling plan, or trucks entering the factory entrance checkpoint and not having a sampling plan, will pass through; and the raw material yard checkpoint refers to the checkpoint that trucks entering the factory entrance checkpoint without a sampling plan or having completed the sampling plan, and trucks without a weighing plan or having completed the weighing plan, will pass through.

[0013] In some optional implementations, the method for establishing the dynamic license plate number database is as follows:

[0014] Upon initial operation, the dynamic license plate number database at each checkpoint is empty;

[0015] According to the management system corresponding to the raw material and fuel procurement contract, obtain the information of the vehicles to be delivered, and add the target license plate number of the vehicles to be delivered to the dynamic license plate number database at the entrance parking lot.

[0016] After successful recognition at the parking lot entrance, the target license plate number is moved from the dynamic license plate number database at the parking lot entrance to the dynamic license plate number database at the parking lot entrance.

[0017] After successful identification at the factory entrance checkpoint, the target license plate number is removed from the dynamic license plate number database at the factory entrance checkpoint. At this point, if there is a sampling plan for the vehicle corresponding to the target license plate number, the target license plate number is added to the dynamic license plate number database at the sampling checkpoint. If there is no sampling plan but there is a weighing plan, the target license plate number is added to the dynamic license plate number database at the weighing checkpoint. If there is neither a sampling plan nor a weighing plan, the target license plate number is added to the dynamic license plate number database at the raw material yard checkpoint.

[0018] After successful identification at the sampling checkpoint, the target license plate number is removed from the dynamic license plate number database of the sampling checkpoint. At this point, if the vehicle corresponding to the target license plate number has a weighing plan, the target license plate number is added to the dynamic license plate number database of the weighing checkpoint. If there is no weighing plan, the target license plate number is added to the dynamic license plate number database of the raw material yard checkpoint.

[0019] After successful recognition at the weighing checkpoint, the target vehicle license plate number is moved from the dynamic license plate number database of the weighing checkpoint to the dynamic license plate number database of the raw material yard checkpoint.

[0020] After successful identification at the raw material yard checkpoint, the target license plate number will be removed from the dynamic license plate number database at the raw material yard checkpoint.

[0021] In some optional implementations, the checkpoint unit is used to acquire the license plate number identified by the main recognition algorithm of the main capture camera. If the license plate number completely matches a license plate in the dynamic license plate number database of the corresponding checkpoint, the recognition process is completed; otherwise, it acquires all license plate numbers identified by each recognition algorithm of each capture camera corresponding to the corresponding checkpoint, performs a complete match on each license plate number in the license plate number database corresponding to the corresponding checkpoint, and selects the successfully matched license plate number if a matching license plate number exists. If all license plate numbers cannot be matched, fuzzy matching is used to determine the final license plate number.

[0022] In some alternative implementations, fuzzy matching includes province fuzziness and specified fuzziness. Province fuzziness means that the province is not considered, and only the letters and numbers of the license plate other than the province are compared. If all characters except the province are completely identical, the match is successful. Specified fuzziness means that a certain character is fuzzily identified as other characters according to the system preset.

[0023] In some optional implementations, the vehicle management module is used to record the results of fuzzy recognition. When the number of fuzzy recognitions meets a preset threshold, the fuzzy matching rules are recorded in the system, and the mapping is automatically performed according to the recorded fuzzy matching rules during the next recognition.

[0024] According to another aspect of the present invention, a method for improving the license plate recognition rate of raw material trucks is provided, comprising:

[0025] (1) Generate a dynamic license plate number database for each type of checkpoint;

[0026] (2) Obtain the license plate number recognized by the camera and compare it with the dynamic license plate number database corresponding to the checkpoint to complete the recognition process;

[0027] (3) The recognition results are self-learned and used in the next recognition process.

[0028] In some alternative implementations, step (1) includes:

[0029] (1.1) Upon initial operation, the dynamic license plate number database at each checkpoint is empty;

[0030] (1.2) According to the management system corresponding to the raw material and fuel procurement contract, obtain the information of the vehicles to be delivered, and add the target license plate number of the vehicles to be delivered to the dynamic license plate number database of the parking lot entrance.

[0031] (1.3) After the vehicle license plate number is successfully identified at the entrance parking lot checkpoint, the target license plate number is moved from the dynamic license plate number database of the entrance parking lot checkpoint to the dynamic license plate number database of the entrance parking lot checkpoint.

[0032] (1.4) After the entry checkpoint is successfully identified, the target license plate number is removed from the dynamic license plate number database of the entry checkpoint. At this time, if the vehicle corresponding to the target license plate number has a sampling plan, the target license plate number is added to the dynamic license plate number database of the sampling checkpoint. If there is no sampling plan but there is a weighing plan, the target license plate number is added to the dynamic license plate number database of the weighing checkpoint. If there is neither a sampling plan nor a weighing plan, the target license plate number is added to the dynamic license plate number database of the raw material yard checkpoint.

[0033] (1.5) After the sampling checkpoint is successfully identified, the target license plate number is removed from the sampling checkpoint dynamic license plate number database. At this time, if the vehicle corresponding to the target license plate number has a weighing plan, the target license plate number is added to the weighing checkpoint dynamic license plate number database. If there is no weighing plan, the target license plate number is added to the raw material yard checkpoint dynamic license plate number database.

[0034] (1.6) After the weighing checkpoint is successfully identified, the target license plate number is moved from the dynamic license plate number database of the weighing checkpoint to the dynamic license plate number database of the raw material yard checkpoint.

[0035] (1.7) After the raw material yard checkpoint is successfully identified, the target license plate number is removed from the dynamic license plate number database of the raw material yard checkpoint.

[0036] In some alternative implementations, step (2) includes:

[0037] (2.1) Obtain the license plate number identified by the main recognition algorithm of the main capture camera. If the license plate number completely matches a license plate in the dynamic license plate number database of the corresponding checkpoint, the recognition process is complete.

[0038] (2.2) If the license plate number does not completely match a certain license plate in the dynamic license plate number database of the corresponding checkpoint, then obtain all the license plate numbers identified by each recognition algorithm of each capture camera corresponding to the corresponding checkpoint, and perform a complete match for each license plate number in the license plate number database corresponding to the corresponding checkpoint. If there is a successfully matched license plate number, then select the successfully matched license plate number. If none of the license plate numbers can be matched, then use fuzzy matching to determine the final license plate number.

[0039] In some alternative implementations, step (3) includes:

[0040] The results of fuzzy recognition are recorded. When the number of fuzzy recognitions meets the preset threshold, the fuzzy matching rules are recorded in the system and automatically mapped according to the recorded fuzzy matching rules in the next recognition.

[0041] In summary, compared with the prior art, the above-described technical solutions conceived by this invention can achieve the following beneficial effects:

[0042] A checkpoint unit with multiple capture cameras and multiple recognition algorithms is proposed, theoretically resulting in a significant improvement in recognition rate. A comprehensive recognition system integrating the hybrid checkpoint unit and vehicle management module is also proposed; the recognition rate can be further improved through fuzzy recognition by the vehicle management module. Furthermore, the vehicle management module's recognition process employs self-learning to automatically correct some recognition errors. Attached Figure Description

[0043] Figure 1 This is a schematic diagram of a system for improving the license plate recognition rate of raw material trucks provided in an embodiment of the present invention;

[0044] Figure 2 This is a schematic diagram of a bayonet provided in an embodiment of the present invention. Detailed Implementation

[0045] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0046] like Figure 1 As shown, the present invention provides a system for improving the license plate recognition rate of raw material and fuel trucks in steel enterprises, comprising two parts: a checkpoint unit and a vehicle management module; for a specific steel plant, it includes multiple checkpoint units and a unified vehicle management module.

[0047] Each checkpoint unit includes one main capture camera and zero to multiple auxiliary capture cameras. Each capture camera includes one main recognition algorithm and zero to multiple auxiliary recognition algorithms.

[0048] The vehicle management module is used to receive and verify the results of license plate recognition.

[0049] Traditional checkpoint units consist of only one camera. Currently, the theoretical recognition rate in laboratories of various manufacturers, both domestically and internationally, is approximately 99%, while the actual project recognition rate is around 95%. This is far from sufficient for the raw material and fuel management system of steel plants.

[0050] With two cameras connected in parallel, the theoretical recognition rate is:

[0051] 1-(1-95%) 2 =99.75%

[0052] By connecting two cameras in parallel, each camera employs a main path recognition algorithm and a single auxiliary path recognition algorithm. Theoretically, the recognition rate is:

[0053] 1-(1-95%) 4 =99.999375%

[0054] However, in actual projects, the improvement in recognition rate was not that significant.

[0055] The solution is as follows:

[0056] To improve the recognition rate, when selecting main and auxiliary capture cameras, try to choose products from different manufacturers, or different models from the same manufacturer.

[0057] The main and auxiliary capture cameras need to be installed at different angles to improve the recognition range;

[0058] The main path recognition algorithm of the capture camera is provided by the manufacturer, while the auxiliary path recognition algorithm can use a custom recognition algorithm based on deep neural networks, which can reduce the miss rate.

[0059] like Figure 2 As shown, the present invention provides a method for improving the license plate recognition rate of raw material and fuel trucks in steel enterprises. Firstly, the checkpoints are logically divided into several types as follows: parking lot entrance checkpoint; factory entrance checkpoint; sampling checkpoint; weighing checkpoint; raw material yard checkpoint;

[0060] in:

[0061] All raw material and fuel trucks must pass through the checkpoint at the factory parking lot entrance.

[0062] The factory entrance checkpoint is where trucks entering the factory parking lot will pass through.

[0063] Sampling checkpoint: This is the entrance checkpoint for vehicles carrying sampling plans.

[0064] The weighing checkpoint, including the truck with the weighing plan, has completed the sampling plan or entered the factory.

[0065] Trucks that do not have a sampling plan will pass through this checkpoint;

[0066] At the raw material yard checkpoint, or at the factory entrance checkpoint, there is no sampling plan or the sampling plan has been completed and there is no...

[0067] Trucks that have completed or are scheduled to be weighed will arrive at this checkpoint.

[0068] Furthermore, a method for improving the license plate recognition rate of raw fuel trucks in iron and steel enterprises provided by the present invention has the following recognition steps:

[0069] Step S1: The vehicle management module generates a dynamic license plate number library for each type of checkpoint.

[0070] Step S2: The checkpoint unit obtains the license plate number of the main recognition algorithm of the main capture camera. If the comparison is successful using comparison algorithm 1 in the dynamic license plate number library of this checkpoint, the recognition process is completed; otherwise, proceed to step S3.

[0071] Step S3: The checkpoint unit obtains all license plate numbers of each recognition algorithm of each capture camera, and performs comparison using comparison algorithm 2 to determine the final license plate number.

[0072] Step S4: The vehicle management module performs self-learning on the recognition results for the next recognition process.

[0073] Among them, comparison algorithm 1 is: the license plate number completely matches a certain license plate in the dynamic license plate number library.

[0074] Among them, comparison algorithm 2 is: perform a complete match for each license plate number in the license plate number library. If a match is successful, select this license plate number. If all license plates cannot be matched, perform fuzzy matching. The fuzzy methods include: province fuzzy, specified fuzzy.

[0075] Province fuzzy means not considering the province and only comparing the other letters and numbers of the license plate except the province. If all characters except the province are exactly the same, the match is successful.

[0076] [[ID=2s4]]Specified fuzzy means that according to the system preset, a certain character can be fuzzy recognized as other characters. For example, 3 is matched as 8. Y is matched as 1. And so on.

[0077] In step S4, record the results of fuzzy recognition. If a certain threshold is met (for example, more than 10 trucks continuously recognize "8" as "3", or recognize "Qing" as "Xin"), the fuzzy matching rule will be automatically recorded in the system.

[0078] In step S4, record the results of fuzzy recognition and perform automatic mapping next time. (For example, if "license plate A" is fuzzy matched as "license plate B" this time, then when "license plate A" is recognized next time, it is directly considered as "license plate B").

[0079] In step S1, the method for establishing the dynamic license plate number library of the vehicle management module is as follows:

[0080] Step S1.1: When the system is initially run, the dynamic license plate number libraries of each checkpoint are empty;

[0081] Step S1.2: According to the management system corresponding to the original fuel purchase contract, obtain the information of the vehicles to be delivered, and add the target license plate number of the vehicle to be delivered to the dynamic license plate number database of the parking lot entrance.

[0082] Step S1.3: After successful recognition at the parking lot entrance checkpoint, the target license plate number is moved from the parking lot entrance checkpoint dynamic license plate number database to the parking lot entrance checkpoint dynamic license plate number database.

[0083] Step S1.4: After successful identification at the factory entrance checkpoint, the target license plate number is removed from the dynamic license plate number database at the factory entrance checkpoint. At this time, if there is a sampling plan for the vehicle corresponding to the target license plate number, the target license plate number is added to the dynamic license plate number database at the sampling checkpoint. If there is no sampling plan but there is a weighing plan, the target license plate number is added to the dynamic license plate number database at the weighing checkpoint. If there is neither a sampling plan nor a weighing plan, the target license plate number is added to the dynamic license plate number database at the raw material yard checkpoint.

[0084] Step S1.5: After the sampling checkpoint is successfully identified, the target license plate number is removed from the sampling checkpoint dynamic license plate number database. At this time, if the vehicle corresponding to the target license plate number has a weighing plan, the target license plate number is added to the weighing checkpoint dynamic license plate number database. If there is no weighing plan, the target license plate number is added to the raw material yard checkpoint dynamic license plate number database.

[0085] Step S1.6: After successful recognition at the weighing checkpoint, the target license plate number is moved from the dynamic license plate number database of the weighing checkpoint to the dynamic license plate number database of the raw material yard checkpoint;

[0086] Step S1.7: After successful identification at the raw material yard checkpoint, remove the target license plate number from the dynamic license plate number database at the raw material yard checkpoint.

[0087] It should be noted that, depending on the implementation needs, the various steps / components described in this application can be broken down into more steps / components, or two or more steps / components or parts of the operation of steps / components can be combined into new steps / components to achieve the purpose of this invention.

[0088] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A system for improving the license plate recognition rate of raw material trucks, characterized in that, include: Checkpoint unit and vehicle management module; The checkpoint unit includes: a parking lot checkpoint, a factory checkpoint, a sampling checkpoint, a weighing checkpoint, and a raw material yard checkpoint; the parking lot checkpoint represents the checkpoint that all raw material and fuel trucks must pass through; the factory checkpoint represents the checkpoint that trucks in the parking lot will pass through; the sampling checkpoint represents the checkpoint that trucks entering the factory checkpoint and included in the sampling plan will pass through; the weighing checkpoint represents the checkpoint that trucks including those with weighing plans who have completed the sampling plan, or those entering the factory checkpoint but not included in the sampling plan, will pass through; the raw material yard checkpoint represents the checkpoint that trucks entering the factory checkpoint without a sampling plan or those that have completed the sampling plan, and those without a weighing plan or those that have completed the weighing plan, will pass through. Each checkpoint unit includes one main capture camera and zero to more auxiliary capture cameras, and each capture camera includes one main recognition algorithm and zero to more auxiliary recognition algorithms; The vehicle management module is used to generate a dynamic license plate number database for each type of checkpoint unit. The method for establishing the dynamic license plate number database is as follows: Upon initial operation, the dynamic license plate number database at each checkpoint is empty; According to the management system corresponding to the raw material and fuel procurement contract, obtain the information of the vehicles to be delivered, and add the target license plate number of the vehicles to be delivered to the dynamic license plate number database at the entrance parking lot. After successful recognition at the parking lot entrance, the target license plate number is moved from the dynamic license plate number database at the parking lot entrance to the dynamic license plate number database at the parking lot entrance. After successful identification at the factory entrance checkpoint, the target license plate number is removed from the dynamic license plate number database at the factory entrance checkpoint. At this point, if there is a sampling plan for the vehicle corresponding to the target license plate number, the target license plate number is added to the dynamic license plate number database at the sampling checkpoint. If there is no sampling plan but there is a weighing plan, the target license plate number is added to the dynamic license plate number database at the weighing checkpoint. If there is neither a sampling plan nor a weighing plan, the target license plate number is added to the dynamic license plate number database at the raw material yard checkpoint. After successful identification at the sampling checkpoint, the target license plate number is removed from the dynamic license plate number database of the sampling checkpoint. At this point, if the vehicle corresponding to the target license plate number has a weighing plan, the target license plate number is added to the dynamic license plate number database of the weighing checkpoint. If there is no weighing plan, the target license plate number is added to the dynamic license plate number database of the raw material yard checkpoint. After successful recognition at the weighing checkpoint, the target vehicle license plate number is moved from the dynamic license plate number database of the weighing checkpoint to the dynamic license plate number database of the raw material yard checkpoint. After successful identification at the raw material yard checkpoint, the target license plate number will be removed from the dynamic license plate number database at the raw material yard checkpoint. The checkpoint unit is used to obtain the license plate number recognized by the main recognition algorithm of the main capture camera. If the license plate number is completely matched with a license plate in the dynamic license plate number database of the corresponding checkpoint, the recognition process is completed; otherwise, it obtains all the license plate numbers recognized by each recognition algorithm of each capture camera corresponding to the corresponding checkpoint, and performs a complete match on each license plate number in the license plate number database corresponding to the corresponding checkpoint. If a successfully matched license plate number exists, the successfully matched license plate number is selected. If all license plate numbers cannot be matched, fuzzy matching is used to determine the final license plate number. The vehicle management module is also used to learn from the recognition results for use in the next recognition process.

2. The system according to claim 1, characterized in that, When selecting the main and auxiliary capture cameras, products from different manufacturers or different models from the same manufacturer are chosen. The installation angles of the main and auxiliary capture cameras differ to improve the recognition range. The main road recognition algorithm of the capture camera uses the manufacturer's own algorithm, while the auxiliary road recognition algorithm uses a custom recognition algorithm based on a deep neural network to reduce the miss rate.

3. The system according to claim 2, characterized in that, Fuzzy matching includes province fuzziness and specified fuzziness. Province fuzziness means that the province is not considered, and only the letters and numbers of the license plate are compared. If all characters except the province are completely identical, the match is successful. Specified fuzziness means that a certain character is fuzzily identified as other characters according to the system preset.

4. The system according to claim 3, characterized in that, The vehicle management module is used to record the results of fuzzy recognition. When the number of fuzzy recognitions meets a preset threshold, the fuzzy matching rules are recorded in the system and automatically mapped according to the recorded fuzzy matching rules during the next recognition.

5. A method for improving the license plate recognition rate of raw material trucks, characterized in that, include: (1) Generate a dynamic license plate number database for each type of checkpoint, the checkpoints including: factory parking lot checkpoint, factory checkpoint, sampling checkpoint, weighing checkpoint and raw material yard checkpoint; the factory parking lot checkpoint represents the checkpoint that all raw material and fuel trucks must pass through; the factory checkpoint represents the checkpoint that trucks in the factory parking lot will pass through; the sampling checkpoint represents the checkpoint that trucks entering the factory checkpoint and having a sampling plan will pass through; the weighing checkpoint represents the checkpoint that trucks with a weighing plan that have completed the sampling plan, or trucks entering the factory checkpoint and not having a sampling plan will pass through; the raw material yard checkpoint represents the checkpoint that trucks entering the factory checkpoint without a sampling plan or having completed the sampling plan, and trucks without a weighing plan or having completed the weighing plan will pass through. (2) Obtain the license plate number recognized by the camera and compare it with the dynamic license plate number database corresponding to the checkpoint to complete the recognition process; (3) The recognition results are self-learned and used in the next recognition process; Step (1) includes: (1.1) Upon initial operation, the dynamic license plate number database at each checkpoint is empty; (1.2) According to the management system corresponding to the raw material and fuel procurement contract, obtain the information of the vehicles to be delivered, and add the target license plate number of the vehicles to be delivered to the dynamic license plate number database of the parking lot entrance. (1.3) After the parking lot entrance gate is successfully identified, the target license plate number is moved from the parking lot entrance gate dynamic license plate number database to the parking lot entrance gate dynamic license plate number database. (1.4) After the entry checkpoint is successfully identified, the target license plate number is removed from the dynamic license plate number database of the entry checkpoint. At this time, if the vehicle corresponding to the target license plate number has a sampling plan, the target license plate number is added to the dynamic license plate number database of the sampling checkpoint. If there is no sampling plan but there is a weighing plan, the target license plate number is added to the dynamic license plate number database of the weighing checkpoint. If there is neither a sampling plan nor a weighing plan, the target license plate number is added to the dynamic license plate number database of the raw material yard checkpoint. (1.5) After the sampling checkpoint is successfully identified, the target license plate number is removed from the sampling checkpoint dynamic license plate number database. At this time, if the vehicle corresponding to the target license plate number has a weighing plan, the target license plate number is added to the weighing checkpoint dynamic license plate number database. If there is no weighing plan, the target license plate number is added to the raw material yard checkpoint dynamic license plate number database. (1.6) After the weighing checkpoint is successfully identified, the target license plate number is moved from the dynamic license plate number database of the weighing checkpoint to the dynamic license plate number database of the raw material yard checkpoint; (1.7) After the raw material yard checkpoint is successfully identified, the target license plate number is removed from the dynamic license plate number database of the raw material yard checkpoint; Step (2) includes: (2.1) Obtain the license plate number identified by the main recognition algorithm of the main capture camera. If the license plate number completely matches a license plate in the dynamic license plate number database of the corresponding checkpoint, the recognition process is complete. (2.2) If the license plate number does not completely match a certain license plate in the dynamic license plate number database of the corresponding checkpoint, then obtain all the license plate numbers identified by each recognition algorithm of each capture camera corresponding to the corresponding checkpoint, and perform a complete match for each license plate number in the license plate number database corresponding to the corresponding checkpoint. If there is a successfully matched license plate number, then select the successfully matched license plate number. If none of the license plate numbers can be matched, then use fuzzy matching to determine the final license plate number.

Images