Inventory vehicle analysis method and device based on video target tracking and license plate recognition

Abstract

The invention provides an inventory vehicle analysis method and device based on video target tracking and license plate recognition, and the method comprises the following steps: capturing vehicle information and tracking a vehicle through employing a multi-target tracking algorithm, and endowing the vehicle with a unique identifier; detecting and identifying the license plate of the tracked uniquely-identified vehicle by adopting a license plate detection and identification algorithm to obtain license plate information of the vehicle, and generating a first record of the stock vehicle; and comparing whether the acquired vehicle information is consistent with the presence vehicle information, updating the presence vehicle information and the first record of the stock vehicles, and obtaining a second record of the stock vehicles. The unique identifier of the vehicle and the license plate number are adopted to filter the repeated license plates, so that the problem that the same license plate of the same vehicle in the video is identified for multiple times is avoided; the unlicensed vehicles can also be distinguished according to the unique identification of the video multi-target tracking algorithm; the number of occupied parking spaces of the parking lot is equal to the number of vehicles in and out of the traffic record; and multiple persons are supported to carry out stocktaking on the same parking lot at the same time.

Description

Technical Field

[0001] The invention belongs to the technical field of vehicle analysis, and in particular relates to a method and a device for analyzing inventory vehicles based on video target tracking and license plate recognition. Background Art

[0002] When counting vehicles in a parking lot, the existing counting method is to manually identify the vehicles in the parking lot and input the vehicle information into the system, which has the following defects:

[0003] Manual inventory requires manual input of information such as license plate numbers, which is inefficient and prone to errors.

[0004] The same vehicle and license plate were identified multiple times in the video, resulting in duplicate records or omissions and underrecording, which led to confusion in the records.

[0005] Unlicensed cars cannot distinguish problems;

[0006] The number of occupied parking lots does not match the number of parking lots calculated based on the entrance and exit records, and there are cases of erroneous records;

[0007] In addition, there is no support for multiple people to inventory the same parking lot at the same time. Summary of the invention

[0008] The purpose of the present invention is to overcome the above problems and provide a method and device for analyzing inventory vehicles based on video target tracking and license plate recognition.

[0009] In order to achieve the above purpose, the technical solution adopted by the present invention is:

[0010] The inventory vehicle analysis method based on video target tracking and license plate recognition includes the following steps:

[0011] A multi-target tracking algorithm is used to capture vehicle information and track vehicles, and a unique identifier is given to the vehicle;

[0012] Using a license plate detection and recognition algorithm to detect and recognize the license plate of the tracked uniquely identified vehicle, obtain the license plate information of the vehicle, and generate the first record of the inventory vehicle;

[0013] Compare the collected vehicle information with the on-site vehicle information to see if they are consistent, update the on-site vehicle information and the first record of the inventory vehicle, and obtain the second record of the inventory vehicle.

[0014] As a further improvement, the license plate detection and recognition algorithm is used to detect and recognize the license plate of the tracked vehicle to obtain the license plate information of the vehicle, including the following steps:

[0015] A vehicle is captured using a multi-target tracking algorithm, and a video frame containing the vehicle is passed to a license plate detection module;

[0016] The license plate detection module adopts a license plate detection algorithm to obtain the license plate position;

[0017] Based on the license plate position, the license plate image is intercepted and passed to the license plate recognition module;

[0018] The license plate recognition module uses a license plate recognition algorithm to obtain the license plate and color of the vehicle;

[0019] in,

[0020] If no license plate exists, the license plate number is unlicensed.

[0021] As a further improvement, the updating of the on-site vehicle information and the first record of the inventory vehicles based on the vehicle request includes the following steps:

[0022] Leave the site or delete redundant vehicle information on the platform, and update the available parking spaces in the parking lot.

[0023] As a further improvement, the updating of the first inventory vehicle record based on the vehicle request further includes the following steps:

[0024] Supplement the missing vehicle information of the parking platform.

[0025] As a further improvement, the updating of the first inventory vehicle record based on the vehicle request further includes the following steps:

[0026] Supplement, leave or delete the information of unlicensed vehicles. Among them, the collected vehicle records that match the parking platform pass records will not be processed.

[0027] As a further improvement, when the current collector finishes collecting vehicle information, the following steps are included:

[0028] If there are other collectors performing collection operations, the current collector is set to the end state;

[0029] If there is no other collector performing collection work, the collection of vehicle information is set to the end state.

[0030] As a further improvement, the multi-target tracking algorithm is used to capture a vehicle, and the video frame containing the vehicle is passed to the license plate detection module, including the following steps:

[0031] A multi-target tracking algorithm is used to capture and track vehicles in the video, a unique identifier is given to the vehicle, and the position of the vehicle in each frame of the video sequence is obtained;

[0032] The position information of the vehicle in each frame of the video sequence is transmitted to the license plate detection module.

[0033] As a further improvement, the step of comparing the collected vehicle information with the on-site vehicle information to see if they are consistent and updating the first record of the inventory vehicle includes the following steps:

[0034] Analyze the parking platform's excess licensed vehicles, missing licensed vehicles, and unlicensed vehicles to be matched;

[0035] Constructing a request for deleting or leaving a vehicle and a vehicle to be recorded again according to the vehicle record information;

[0036] Based on the vehicle request, the on-site vehicle information and the first record of the inventory vehicles are updated.

[0037] On the other hand, the present invention also provides an inventory vehicle analysis system based on video target tracking and license plate recognition, the system comprising at least one processor and a memory storing instructions, which implement the method in the above technical solution when the instructions are executed by at least one processor.

[0038] A computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement the method described in the steps of the above technical solution.

[0039] A computer program product, wherein the computer program product comprises a computer program, and when the computer program is executed by a processor, the method described in the steps of the above technical solution is implemented.

[0040] Beneficial effects of the present invention:

[0041] The present invention utilizes multi-target tracking and license plate recognition technologies in series to realize a complete closed loop of inventory vehicle inspection; it brings convenience to vehicle inventory inspection, makes recording more convenient, improves efficiency, and reduces the error rate; it uses the vehicle unique identification and license plate number to filter repeated license plates, thereby avoiding the problem of the same vehicle and the same license plate being recognized multiple times in the video; at the same time, vehicles without license plates can also be distinguished according to the unique identification of the video multi-target tracking algorithm; by deleting excess vehicle departures in vehicle records, supplementing missing vehicles, and matching records of vehicles without license plates, it is achieved that the number of occupied parking lots and the number of vehicles in and out of the parking lot are equal, thereby improving the accuracy of parking space use; it supports multiple people to conduct inventory inspections in the same parking lot at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The present invention has the following accompanying drawings:

[0043] Figure 1 It is a flowchart of a method for analyzing inventory vehicles based on video target tracking and license plate recognition according to some specific embodiments of the present invention;

[0044] Figure 2It is a structural schematic diagram of a vehicle inventory analysis device based on video target tracking and license plate recognition in some specific embodiments of the present invention;

[0045] Figure 3 is a flow chart of a method for analyzing inventory vehicles based on video target tracking and license plate recognition according to some specific embodiments of the present invention;

[0046] Figure 4 Schematic diagram of the inventory vehicle analysis system based on video target tracking and license plate recognition of the present invention. DETAILED DESCRIPTION

[0047] The following examples further illustrate the content of the present invention, but should not be construed as limiting the present invention. Without departing from the spirit and substance of the present invention, modifications or substitutions made to the methods, steps or conditions of the present invention all fall within the scope of the present invention.

[0048] Embodiment 1:

[0049] like Figure 1 As shown, the present invention provides a method for analyzing inventory vehicles based on video target tracking and license plate recognition, comprising the following steps:

[0050] Step S10: Use a multi-target tracking algorithm to capture vehicle information and track vehicles, and assign a unique identifier to the vehicle. The collector can use a mobile phone APP to take pictures of the vehicle, collect vehicle information, capture and track the vehicle with the help of a multi-target tracking algorithm, and uniquely identify the captured vehicle to provide support for subsequent comparison and avoidance of recording errors. When the collector starts to count the inventory, the database in the parking platform generates and saves the inventory record, recording the operator and vehicle inventory status.

[0051] The present invention uses the vehicle unique identifier and the license plate number to filter repeated license plates, thereby avoiding the problem that the same vehicle and the same license plate are recognized multiple times in the video; vehicles without license plates can also be distinguished according to the unique identifier of the video multi-target tracking algorithm.

[0052] Step S11: using a license plate detection and recognition algorithm to detect and recognize the license plate of the tracked uniquely identified vehicle, obtain the license plate information of the vehicle, and generate the first record of the inventory vehicle.

[0053] For example, the target detection algorithm can be used to detect the location of the license plate; the license plate recognition algorithm can be used to identify the license plate number and license plate color from the captured license plate image. Then, the license plate number, license plate color, area, parking lot and time are uploaded to generate an inventory vehicle record, that is, the first inventory vehicle record is generated.

[0054] Step S12: Compare the collected vehicle information with the on-site vehicle information to see if they are consistent, update the on-site vehicle information and the first record of the inventory vehicles, and obtain the second record of the inventory vehicles.

[0055] For example, redundant vehicle records are deleted, missing vehicle records are supplemented, and pass records without license plates are queried. The erroneous vehicle information is corrected, and then the vacant parking spaces in the parking lot are vacated, the first record of inventory vehicles is updated, and the second record of inventory vehicles is obtained to facilitate future parking of vehicles and improve the accuracy of parking space utilization.

[0056] Step S12 includes the following steps:

[0057] Step S120: analyzing the vehicle record information of excess licensed vehicles, missing licensed vehicles, and unlicensed vehicles to be matched on the parking platform;

[0058] Step S121: constructing a request for deleting or leaving the vehicle and for supplementing the vehicle according to the vehicle record information;

[0059] The present invention achieves equality between the number of occupied parking lots and the number of vehicles in and out of the parking lot by deleting excess vehicle departures from the vehicle records, supplementing missing vehicles, and matching unlicensed vehicle records, thereby improving the accuracy of parking space use.

[0060] Step S122: Based on the vehicle request, update the on-site vehicle information and the first record of the inventory vehicles.

[0061] For example, updating the on-site vehicle information includes vehicle records, occupied parking spaces and free parking spaces, and the first record of inventory vehicles includes record processing status and processing methods.

[0062] Step S11 includes the following steps:

[0063] Step S110: a vehicle is captured using a multi-target tracking algorithm, and a video frame containing the vehicle is transmitted to a license plate detection module;

[0064] Step S111: The license plate detection module uses the public YOLO series target tracking algorithm. First, the license plate position in the image is marked as a label, and the image is transformed into a rotation, translation, fusion, etc. as a data set, which is sent to the vehicle detection algorithm for training, that is, to minimize the loss function. The loss function is as follows

[0065]

[0066] Loss = L box +L cls +L obj

[0067] The loss calculation of the model includes the license plate location loss L box ,License plate classification loss L cls , whether it is the license plate confidence loss L obj, the loss function Loss is the weighted sum of three types of losses, where λ coord is the license plate location loss weight, λ class is the loss weight of the license plate category, λ obj and λ noobj are the weight coefficients of whether it is a license plate or not. Among them, B is the number of targets whose labels are assigned to the prior box, S 2 is the number of grids divided into by this scale, and I is the indicator function of whether there is a license plate in the current prediction frame.

[0068] Train the detection model. Then input the image to be predicted into the detection model to obtain the license plate position of the image to be predicted.

[0069] Step S112: based on the license plate position, intercept the license plate image and transmit it to the license plate recognition module;

[0070] Step S113: The license plate recognition module annotates the captured license plate image, performs enhancement processing such as rotation and blurring on the license plate image, maps the license plate number into a digital sequence according to the characters, and directly maps the license plate color into a number, and makes a data set, which is sent to the convolutional neural network model for training. The license plate number loss function adopts the CTCLoss model.

[0071]

[0072] Using a license plate recognition algorithm to obtain the license plate and color of the vehicle;

[0073] The license plate color loss function is the softmax loss function, and the formula is

[0074]

[0075] Where N represents the number of vehicle samples, that is, the number of vehicle samples in the training set, and C represents the number of license plate colors6. Indicates the label of the nth vehicle sample belonging to category c, and its value is 0 or 1. Represents the predicted value of the nth vehicle sample.

[0076] After training the license plate number and license plate color recognition model, input the license plate screenshot of the image to be predicted to get the license plate number and license plate color. If there is no license plate, the license plate number is unlicensed. During the collection process, unlicensed vehicles can also be identified and then marked as unlicensed.

[0077] Step S122 includes the following steps:

[0078] Step S1220: Leave the parking lot or delete the redundant vehicle information on the platform, and update the available parking spaces in the parking lot.

[0079] Step S122 further includes the following steps:

[0080] Step S1221: Supplement the missing vehicle information of the parking platform. During the recognition process, some licensed vehicles are not recognized. After comparison, they need to be supplemented so that the vehicle parking information can be more accurate.

[0081] Step S122 further includes the following steps:

[0082] Step S1222: Supplement, leave or delete the information of unlicensed vehicles, among which, the collected vehicle records that match the parking platform passage records will not be processed.

[0083] In some embodiments of the present invention, when the current collector finishes collecting vehicle information, multiple persons are supported to conduct inventory of the same parking lot at the same time, including:

[0084] If there are other collectors performing collection operations, the current collector is set to the end state;

[0085] If there is no other collector performing collection work, the collection of vehicle information is set to the end state.

[0086] For example, at the end of the inventory, the collection vehicle record and the status of the collector (inspector) must be updated in the database.

[0087] Step S110 includes the following steps:

[0088] Step S1100: using a multi-target tracking algorithm to capture and track the vehicle in the video, assigning a unique identifier to the vehicle, and obtaining the position of the vehicle in each frame of the video sequence;

[0089] Step S1101: The position information of the vehicle in each frame of the video sequence is transmitted to a license plate detection module.

[0090] For example, by directly transmitting the position of the vehicle in the video to the license plate detection module, the license plate can be found directly from the position in the data frame more quickly, and then the license plate can be identified, which is faster than directly feeding the video without marking it for searching.

[0091] Embodiment 2:

[0092] The present invention also provides a vehicle inventory analysis device based on video target tracking and license plate recognition, comprising:

[0093] A tracking module, used to capture vehicle information and track vehicles using a multi-target tracking algorithm, and to assign a unique identifier to the vehicle;

[0094] The acquisition module is used to detect and identify the license plate of the tracked uniquely identified vehicle using a license plate detection and recognition algorithm, obtain the license plate information of the vehicle, and generate the first record of the inventory vehicle;

[0095] A comparison module is used to compare the collected vehicle information with the on-site vehicle information to see if they are consistent, update the on-site vehicle information and the first record of the inventory vehicle, and obtain the second record of the inventory vehicle;

[0096] The control module includes:

[0097] An analysis module is used to analyze the vehicle record information of excess licensed vehicles, missing licensed vehicles, and unlicensed vehicles to be matched on the parking platform;

[0098] A construction module, used for constructing a request for deleting or leaving a vehicle and a request for supplementary recording according to the vehicle record information;

[0099] The updating module is used for updating the on-site vehicle information and the first record of the inventory vehicles based on the vehicle request.

[0100] In some embodiments of the present invention, the acquisition module includes:

[0101] A first transmission module is used to capture a vehicle using a multi-target tracking algorithm and transmit a video frame containing the vehicle to a license plate detection module;

[0102] A first calculation module, used for the license plate detection module to use a license plate detection algorithm to obtain a license plate position;

[0103] The second transmission module is used to intercept the license plate image based on the license plate position and transmit it to the license plate recognition module;

[0104] The second calculation module is used for the license plate recognition module to use the license plate recognition algorithm to obtain the license plate and color of the vehicle; wherein, if the license plate does not exist, the license plate number is unlicensed.

[0105] In some embodiments of the present invention, the update module includes:

[0106] The redundant vehicle processing module is used to leave or delete the redundant vehicle information on the platform and update the vacant parking spaces in the parking lot.

[0107] In some embodiments of the present invention, the update module further includes:

[0108] The missing vehicle processing module is used to supplement the missing vehicle information of the parking platform.

[0109] In some embodiments of the present invention, the update module further includes:

[0110] The unlicensed vehicle processing module is used to supplement, leave or delete the unlicensed vehicle information. Among them, the collected vehicle records that match the parking platform pass records will not be processed.

[0111] In some embodiments of the present invention, when the current collector finishes collecting vehicle information, the following steps are included:

[0112] If there are other collectors performing collection operations, the current collector is set to the end state;

[0113] If there is no other collector performing collection work, the collection of vehicle information is set to the end state.

[0114] In some embodiments of the present invention, the first transmission module includes:

[0115] A position identification module is used to capture and track vehicles in the video using a multi-target tracking algorithm, assign a unique identifier to the vehicle, and obtain the position of the vehicle in each frame of the video sequence;

[0116] The third transmission module is used to transmit the position information of the vehicle in each frame of the video sequence to the license plate detection module.

[0117] Embodiment 3:

[0118] like Figure 3 As shown, the inventory vehicle analysis method and device based on video target tracking and license plate recognition include the following technical contents:

[0119] Step S1: The inspector shoots a video through the mobile phone app. Specifically, the inspector opens the inventory check app, clicks on start inventory check, and uses the mobile phone to shoot the vehicles in the parking lot. The background program creates an inventory check record for the parking lot and records the inspector's inventory check status in the parking lot as inventory check in progress.

[0120] Step S2: The multi-target tracking algorithm captures and tracks the vehicle in the video, assigns a unique identifier to the vehicle, and gives the position of the vehicle in each frame of the video sequence. The multi-target tracking algorithm is an existing deepsort algorithm, and the model file is trained on its own data set.

[0121] Step S3: The license plate recognition algorithm detects and recognizes the license plate in the video frame. Specifically, the multi-target tracking algorithm captures a new target, passes the video frame containing the target to the license plate detection algorithm, the license plate detection module gives the license plate position, captures the license plate image, and passes it to the license plate recognition module, which gives the license plate and color of the vehicle. If there is no license plate, the license plate number is unlicensed.

[0122] Step S4: The app uploads the inventory records to the background and saves them to the database. Specifically, the app reports the current time, parking lot, parking area, recognized license plate number, license plate color, vehicle image, license plate image, etc. to the background program, and the program saves the inventory records to the database for easy operation and viewing.

[0123] Step S5: Complete the inventory by taking photos of all vehicles in the current parking lot. Specifically, click the End Inventory button. If there are other inspectors in the parking lot, set the current inspector's status to Ended and return to the list of inspectors currently inventorying. If there are no other inspectors in the inventory, the inventory is completed.

[0124] Step S6: The app background program will compare the on-site vehicles obtained from the inventory with the on-site vehicle records, and display and process the records of excess licensed vehicles, missing licensed vehicles, and unlicensed vehicles to be matched on the parking platform.

[0125] Step S7: The inspector processes the inventory results, specifically: Licensed vehicle records: For those missing from the parking platform, they will be supplemented according to the inventory information (including the time the picture was taken, the parking area, the license plate number and color), and the redundant ones on the platform will be removed or deleted, and the parking lot's vacant spaces will be updated. For unlicensed vehicle records, the app will query and list all the unlicensed vehicle traffic records on the parking platform, and the inspector can supplement, remove or delete them.

[0126] Through the method and device, the present invention utilizes multi-target tracking and license plate recognition technologies in series to realize a complete closed loop of inventory vehicle inspection; it brings convenience to vehicle inventory inspection, makes recording more convenient, improves efficiency, and reduces the error rate; uses the vehicle unique identification and license plate number to filter repeated license plates, avoiding the problem of the same vehicle and the same license plate being recognized multiple times in the video; at the same time, unlicensed vehicles can also be distinguished according to the unique identification of the video multi-target tracking algorithm; by deleting the excess vehicle departures in the vehicle records, supplementing the missing vehicles, and matching the unlicensed vehicle records, the number of occupied parking lots and the number of vehicles in the entry and exit records are equal, thereby improving the accuracy of parking space use; it supports multiple people to inventory the same parking lot at the same time.

[0127] In some specific embodiments, Figure 4 As shown, a vehicle inventory analysis system based on video target tracking and license plate recognition includes at least one processor and a memory storing instructions to implement any vehicle inventory analysis method based on video target tracking and license plate recognition in the above technical solutions.

[0128] In some specific embodiments, a computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement the steps in the above technical solution.

[0129] In some specific embodiments, a computer program product includes a computer program, and when the computer program is executed by a processor, the method of implementing the steps in the above technical solution.

[0130] The embodiments and functional operations of the subject matter described in this specification may be implemented in: digital electronic circuitry, tangibly implemented computer software or firmware, computer hardware, including the structures disclosed in this specification and their structural equivalents, or a combination of more than one of the above. The embodiments of the subject matter described in this specification may be implemented as one or more computer programs, i.e., one or more modules of computer program instructions encoded on one or more tangible non-transitory program carriers, for execution by a data processing device or to control the operation of a data processing device.

[0131] Alternatively or additionally, the program instructions may be encoded on an artificially generated propagation signal, such as a machine-generated electrical, optical, or electromagnetic signal, which is generated to encode information for transmission to an appropriate receiver device for execution by a data processing device. The computer storage medium may be a machine-readable storage device, a machine-readable storage substrate, a random or serial access memory device, or a combination of one or more of the above devices.

[0132] A computer program (which may also be referred to or described as a program, software, software application, module, software module, script or code) may be written in any form of programming language, including compiled or interpreted languages ​​or declarative or procedural languages, and the computer program may be deployed in any form, including as a stand-alone program or as a module, component, subroutine or other unit suitable for use in a computing environment. A computer program may, but need not, correspond to a file in a file system. A program may be stored in a portion of a file that stores other programs or data, for example, one or more scripts stored in the following: in a markup language document; in a single file dedicated to a related program; or in multiple collaborative files, for example, files storing one or more modules, subroutines or code portions. A computer program may be deployed to execute on one or more computers, the computers being located in one place or distributed to multiple locations and interconnected by a communication network.

[0133] The contents not described in detail in this specification belong to the prior art known to professional and technical personnel in this field.

[0134] Although the present invention has been described in detail above by general description, specific implementation methods and experiments, it is obvious to those skilled in the art that some modifications or improvements can be made to the present invention. Therefore, these modifications or improvements made without departing from the spirit of the present invention all belong to the scope of protection claimed by the present invention.

Claims

1. A method for analyzing inventory vehicles based on video target tracking and license plate recognition, characterized in that: The following steps are involved: A multi-target tracking algorithm is used to capture vehicle information and track vehicles, and a unique identifier is given to the vehicle; Using a license plate detection and recognition algorithm to detect and recognize the license plate of the tracked uniquely identified vehicle, obtain the license plate information of the vehicle, and generate the first record of the inventory vehicle; Compare the collected vehicle information with the on-site vehicle information to see if they are consistent, update the on-site vehicle information and the first record of the inventory vehicle, and obtain the second record of the inventory vehicle and accurate on-site vehicle information.

2. The method for analyzing inventory vehicles based on video target tracking and license plate recognition according to claim 1 is characterized in that: The method of using a license plate detection and recognition algorithm to detect and recognize the license plate of the tracked vehicle to obtain the license plate information of the vehicle includes the following steps: A vehicle is captured using a multi-target tracking algorithm, and a video frame containing the vehicle is passed to a license plate detection module; The license plate detection module adopts a license plate detection algorithm to obtain the license plate position; Based on the license plate position, the license plate image is intercepted and passed to the license plate recognition module; The license plate recognition module uses a license plate recognition algorithm to obtain the license plate and color of the vehicle; wherein, If no license plate exists, the license plate number is unlicensed.

3. The method for analyzing inventory vehicles based on video target tracking and license plate recognition according to claim 1 is characterized in that: The updating of the on-site vehicle information and the first record of the inventory vehicle based on the vehicle request comprises the following steps: Leave the site or delete redundant vehicle information on the platform, and update the available parking spaces in the parking lot.

4. The method for analyzing inventory vehicles based on video target tracking and license plate recognition according to claim 1 is characterized in that: The updating of the first inventory vehicle record based on the vehicle request further includes the following steps: Supplement the missing vehicle information of the parking platform.

5. The method for analyzing inventory vehicles based on video target tracking and license plate recognition according to claim 1 is characterized in that: The updating of the on-site vehicle information and the first record of the inventory vehicle based on the vehicle request further includes the following steps: Supplement, leave or delete the information of unlicensed vehicles. Among them, the collected vehicle records that match the parking platform pass records will not be processed.

6. The method for analyzing inventory vehicles based on video target tracking and license plate recognition according to claim 1, characterized in that: When the current collector finishes collecting vehicle information, the following steps are included: If there are other collectors performing collection operations, the current collector is set to the end state; If there is no other collector performing collection work, the collection of vehicle information is set to the end state.

7. The method for analyzing inventory vehicles based on video target tracking and license plate recognition according to claim 2 is characterized in that: The method of capturing a vehicle using a multi-target tracking algorithm and transmitting a video frame containing the vehicle to a license plate detection module comprises the following steps: A multi-target tracking algorithm is used to capture and track vehicles in the video, a unique identifier is given to the vehicle, and the position of the vehicle in each frame of the video sequence is obtained; The position information of the vehicle in each frame of the video sequence is transmitted to the license plate detection module.

8. The method for analyzing inventory vehicles based on video target tracking and license plate recognition according to claim 1, characterized in that: The step of comparing the collected vehicle information and the on-site vehicle information to see whether they are consistent, and updating the on-site vehicle information and the first record of the inventory vehicle, comprises the following steps: Analyze the parking platform's excess licensed vehicles, missing licensed vehicles, and unlicensed vehicles to be matched; Constructing a request for deleting or leaving a vehicle and a vehicle to be recorded again according to the vehicle record information; Based on the vehicle request, the on-site vehicle information and the first record of the inventory vehicles are updated.

9. A vehicle inventory analysis system based on video target tracking and license plate recognition, the system comprising at least one processor and a memory storing instructions, which, when executed by at least one processor, implements the vehicle inventory analysis method based on video target tracking and license plate recognition as described in any one of claims 1 to 8.

Images