Inspection vehicle berth positioning system and method based on deep learning

Abstract

The invention discloses an inspection vehicle berth positioning system and method based on deep learning, and relates to the technical field of intelligent traffic. The system comprises an image acquisition module which is designed to acquire a parking space image through a camera; the vehicle positioning module is designed to obtain positioning information of the video inspection vehicle through GPS positioning equipment; the data processing module is designed to preprocess the parking space image and analyze image data according to an image algorithm; and the mobile communication module is designed to transmit license plate and parking space information to an operation management platform through a mobile communication technology, and a deep learning scheme is added to improve the recognition precision and speed of the OCR character recognition algorithm. According to the technical scheme of the invention, the method can achieve the quick recognition of the license plate information of the vehicle and the matching of the parking space in a parking charging scene, and is higher in application value.

Description

[0001] This application is a divisional application of a Chinese invention patent application with an application date of "2024.09.26", an application number of "202411349564.6", and an invention title of "An OCR-based Patrol Vehicle Berth Positioning System and Method". Technical Field

[0002] The present invention relates to the field of intelligent transportation technology, and particularly to an OCR-based patrol vehicle berth positioning system and method. Background Art

[0003] In recent years, the number of motor vehicles in China has been continuously increasing, and cars have become an essential means of transportation for people's daily travel. The problem of parking difficulty has become increasingly prominent. Especially near busy commercial areas, residential communities, and public transportation hubs, finding a suitable parking space has become a headache for many drivers. In order to improve the utilization rate of on-street berths, many road surface intelligent patrol vehicles have emerged on the market. These patrol vehicles improve the efficiency of road surface patrol while reducing manual participation. Using a video patrol vehicle to collect the parking conditions of on-street parking and taking pictures of the vehicles on the berths, among which, matching the vehicle to the corresponding berth is one of the core functions of the video patrol vehicle and an important part affecting its patrol accuracy. Currently, most existing berth matching schemes rely on high-frequency and high-precision GPS devices. According to the real-time GPS data received by the GPS devices, the GPS longitude and latitude static data of each berth are pre-calibrated to complete the berth matching work. However, this scheme still has some significant problems: before the video patrol vehicle is put into use, it is necessary to collect the longitude and latitude data of the starting and ending points of each berth. The workload of the collection process is extremely heavy, and the amount of data to be collected is extremely large; in an environment with weak satellite positioning signals, such as being blocked by high-rise buildings, trees, or signal shielding, inaccurate GPS positioning will cause a deviation between the pre-calibrated longitude and latitude data of the berth and the actual longitude and latitude data of the berth, thus affecting the normal operation of the video patrol vehicle. Therefore, the research and development of a simple and efficient patrol vehicle berth positioning method is extremely urgent.

[0004] Currently, most existing patrol vehicle berth positioning methods have defects in one aspect or another. For example, the scheme of NB geomagnetism combined with a video patrol vehicle, although it has the advantages of convenient device layout, high precision, and low false alarm rate, the performance stability of the dual-mode geomagnetism is insufficient in special weather environments, while the triple-mode geomagnetism has a relatively high cost; the SLAM scheme, although it can estimate the pose of a moving vehicle in real time and construct a map of the surrounding environment, visual SLAM cannot work in dark or textureless areas, and the application scenario of laser SLAM is relatively limited. Therefore, based on this difficult problem, this application proposes an OCR-based patrol vehicle berth positioning system and method. Summary of the Invention

[0005] Technical objective To solve the above problems, the objective of the present invention is to provide an OCR-based inspection vehicle berth positioning system and method, which not only solves the problem of time-consuming and laborious manual collection of berth longitude and latitude data in the traditional berth positioning process, but also avoids the occurrence of inaccurate longitude and latitude data caused by human factors, realizes the simplification and automation of berth positioning, significantly improves the inspection accuracy of the video inspection vehicle and its adaptability to different scenarios, and avoids the application obstacle problem of the video inspection vehicle caused by objective factors such as satellite positioning.

[0006] Technical solution To achieve the above objective, the present invention provides an OCR-based inspection vehicle berth positioning system and method, which identifies the berth numbers sprayed on the ground through image segmentation and OCR text recognition algorithms, so as to achieve precise matching of on-street berths.

[0007] In the first aspect, the present invention provides an OCR-based inspection vehicle berth positioning system, including: An image acquisition module, which acquires berth images through a camera, wherein the berth images include the license plate number of the vehicle and the berth number; A vehicle positioning module, which obtains the positioning information of the video inspection vehicle through a GPS positioning device; A data processing module, which preprocesses the berth images and analyzes the image data according to image algorithms to obtain license plate information and berth information, and determines the berth corresponding to the captured image based on the time of the video frame, the time corresponding to the license plate capture, and the position of the berth number in the image; wherein the image algorithms include OCR text recognition algorithms; A mobile communication module, which transmits the license plate and parking space information to the operation and management platform through mobile communication technology.

[0008] Furthermore, the license plate number of the vehicle is obtained by real-time shooting and recognition through an intelligent license plate recognition camera installed on the vehicle inspection vehicle; the berth number is obtained by real-time shooting and recognition through a special camera installed on the vehicle inspection vehicle for collecting clear berth number video streams; the intelligent license plate recognition camera and the special camera are installed at the front end of the video inspection vehicle according to the relative positions of the license plate of the vehicle and the berth number.

[0009] Furthermore, after receiving the video stream collected by the berth number acquisition camera, the data processing module obtains the berth number data through the built-in OCR character recognition algorithm of the data processing module. It preprocesses the frame in the video stream, segments the berth number characters in the frame, and recognizes the content of the berth number characters, thereby obtaining the berth number data. It not only solves the problem of time-consuming and laborious manual collection of berth longitude and latitude data in the traditional berth positioning process, but also avoids the occurrence of inaccurate longitude and latitude data caused by human factors, realizes the simplification and automation of berth positioning, significantly improves the inspection accuracy of the video inspection vehicle and its adaptability to different scenarios, and avoids the application obstacle problem of the video inspection vehicle caused by objective factors such as satellite positioning.

[0010] Furthermore, the preprocessing of the frame in the video stream by the data processing module specifically includes: marking the text area, determining the position of the text in each document by marking the text area of the collected sample data, thereby constructing a model to learn the spatial structure of the text and improving the recognition accuracy of the text; extracting the text content, extracting the text content in the marked text area and establishing label data corresponding to the text content for supervised learning; data cleaning and preprocessing, cleaning and preprocessing the extracted text content, including removing noise, adjusting the size and direction, etc., to ensure the quality of the text data.

[0011] Furthermore, the OCR character recognition algorithm supports multiple models, including traditional template matching algorithms, support vector machines, Bayesian classification algorithms, and neural network models based on deep learning, etc. It can select a suitable model for training according to specific application scenarios and requirements; after the model training is completed, the data processing module improves the accuracy of the OCR character recognition algorithm through dataset optimization, feature selection, and model structure optimization. Among them, dataset optimization is specifically to use a more accurate and comprehensive dataset for training, and optimize the dataset by increasing the sample size, deleting duplicate samples, increasing randomness, etc.; feature selection is specifically to select the most important features for the recognition task and reduce unnecessary features, thereby accelerating the response speed of the OCR system. The selection of relevant features is based on feature evaluation methods, such as information gain, correlation coefficient, etc.

[0012] Furthermore, the data processing module improves the recognition accuracy and speed of the OCR character recognition algorithm based on a deep learning solution. It uses a convolutional neural network as the basic model and continuously improves the recognition ability of the OCR character recognition algorithm by increasing the amount of training data. The deep learning solution includes steps such as data preprocessing, constructing a convolutional neural network model, calculating the loss function, backpropagation and optimization, and regularization. It can help the OCR system better process characters in complex environments, improve the recognition accuracy of the OCR system for characters under non-ideal conditions, significantly speed up the operation speed of the model, and reduce the need for manual feature engineering by automatically learning features in the image, thus simplifying the development process.

[0013] Furthermore, the spraying of the berth number is based on a custom number spraying standard dedicated to the OCR system. The number spraying standard involves the font, font size, and color of the berth number, so as to improve the applicability of the OCR algorithm for berth numbers. It helps the OCR system more accurately identify the berth number, reduces the time required for the OCR algorithm to process images, enables the OCR system to more easily adapt to different environmental conditions, and improves the recognition ability of the OCR system in diverse scenarios.

[0014] Furthermore, the data processing module determines whether the berth information of this berth is correct by comprehensively comparing the berth information of the previous berth and the positioning information of the video inspection vehicle with the berth information of this berth and the positioning information of the video inspection vehicle. For example, if the berth number of the previous berth is 00050 and the physical distance between the positioning of the video inspection vehicle at the previous berth and the positioning of the video inspection vehicle at this berth is only 6 meters, but the calculated berth number of this berth is 00031, then it is judged that there may be a situation where the berth number of this berth is not correctly recognized. At this time, comprehensive verification can be carried out according to the berth information and positioning information of the next berth. If the video inspection vehicle only moves about 6 meters and then recognizes the berth number of the next berth as 00052, then it can be basically determined that the actual berth number of this berth should be 00051. The reason for the incorrect recognition of the berth number of this berth may be that the berth number is blocked, contaminated, or damaged. Similar abnormal situations can be submitted to the operation management platform for subsequent processing, assisting the image algorithm to judge the license plate information and berth information from multiple angles, improving the recognition and positioning accuracy, and improving the operation management efficiency.

[0015] Further, the mobile communication module establishes a parking space geographic information database based on a geographic information scheme, which displays the location distribution map of parking berths in the form of a map, including: matching the data recognized by the OCR system with the geographic information in the geographic information scheme database, and the geographic information scheme database should contain information such as the geographic coordinates, attribution area, and type of the berth; combining the real-time collected data with the geographic information scheme, and performing spatial analysis and processing, and using the geographic information scheme to monitor and manage the vehicle position and berth status in real time; displaying the berth status, vehicle position, and relevant statistical information through the map view. It can quickly respond to parking problems by monitoring the status of on-street berths and the flow of nearby vehicles in real time, can more effectively allocate and manage berth resources, provide fast and accurate berth information for drivers, and reduce the time for drivers to search for berths.

[0016] In a second aspect, the present invention also provides an OCR-based inspection vehicle berth positioning method, which is based on the system described in the first aspect above, and includes: Obtaining a berth image through a camera installed on the video inspection vehicle; Obtaining the positioning information of the video inspection vehicle through a GPS positioning device; Preprocessing the berth image and analyzing the image data according to the image algorithm; Outputting the license plate and parking space information through mobile communication technology.

[0017] In a third aspect, the present invention also provides a computer-readable storage medium, in which a computer program is stored, and when the computer program is executed by a processor, it implements the aforementioned OCR-based inspection vehicle berth positioning method.

[0018] The present invention recognizes the berth numbers sprayed on the ground through image segmentation and OCR character recognition algorithms, and the spraying of the berth numbers is based on a custom number spraying standard for the OCR system; and establishes a parking space geographic information database based on a geographic information scheme, and displays the location distribution map of parking berths in the form of a map, and improves the recognition accuracy and speed of the OCR character recognition algorithm based on a deep learning scheme. The system and method not only solve the problem of time-consuming and laborious manual collection of berth longitude and latitude data in the traditional berth positioning process, but also avoid the occurrence of inaccurate longitude and latitude data caused by human factors, realize the simplification and automation of berth positioning, significantly improve the inspection accuracy of the video inspection vehicle and its adaptability to different scenarios, and avoid the application obstacle problem of the video inspection vehicle caused by objective factors such as satellite positioning.

[0019] Beneficial effects By implementing the above-mentioned OCR-based inspection vehicle berth positioning system and method provided by the present invention, the following technical effects are achieved: (1) The present invention identifies the berth numbers sprayed on the ground through image segmentation and OCR text recognition algorithms. It not only solves the problem of time-consuming and laborious manual collection of berth longitude and latitude data in the traditional berth positioning process, but also avoids the occurrence of inaccurate longitude and latitude data caused by human factors, realizes the simplification and automation of berth positioning, significantly improves the inspection accuracy of the video inspection vehicle and its adaptability to different scenarios, and avoids the application obstacle problems of the video inspection vehicle caused by objective factors such as satellite positioning.

[0020] (2) The spraying of the berth numbers in the present invention is based on the number spraying standard dedicated to the custom OCR system. It helps the OCR system to more accurately identify the berth numbers, reduces the time required for the OCR algorithm to process images, enables the OCR system to more easily adapt to different environmental conditions, and improves the recognition ability of the OCR system in diverse scenarios.

[0021] (3) The present invention establishes a parking space geographic information database based on a geographic information scheme and displays the location distribution map of parking berths in the form of a map. It can quickly respond to parking problems by real-time monitoring the status of on-street berths and the flow of nearby vehicles, can more effectively allocate and manage berth resources, provide fast and accurate berth information for drivers, and reduce the time for drivers to search for berths.

[0022] (4) The present invention improves the recognition accuracy and speed of the OCR text recognition algorithm based on a deep learning scheme and uses a convolutional neural network as the basic model. It can help the OCR system better process text in complex environments, improve the recognition accuracy of the OCR system for text under non-ideal conditions, significantly speed up the operation speed of the model, and reduce the need for manual feature engineering by automatically learning features in images, thus simplifying the development process. Brief Description of the Drawings

[0023] To make the above-mentioned OCR-based inspection vehicle berth positioning system and method of the present invention more obvious and understandable, the following will briefly introduce the drawings required for the specific implementation of the present invention. Obviously, the drawings in the following description are only some embodiments of the present invention, and those skilled in the art can obtain other drawings according to these drawings without creative efforts.

[0024] Figure 1 Indicates the schematic diagram of the principle of the inspection vehicle berth positioning system; Figure 2 Indicates the schematic diagram of the position of the berth number acquisition camera; Figure 3 Indicates the schematic diagram of the berth number spraying standard; Figure 4 Indicates the schematic diagram of the effect of the OCR text recognition algorithm; Figure 5 Schematic diagram showing the effect of the geographic information solution. Specific implementation mode

[0025] Example 1: A PTK-based inspection vehicle berth positioning system is provided. As Figure 1 shown, the inspection vehicle is equipped with a GPS receiver to receive signals from satellites, and these signals include time stamps and the position information of the satellites; the reference station is a fixed GPS receiver, and its position is known. The reference station receives the same satellite signals and broadcasts the received data to the inspection vehicle through a radio; the radio, as a communication medium, transmits the correction information received by the reference station to the inspection vehicle in real time; after the PTK positioning module on the inspection vehicle receives the correction information broadcast by the reference station, it uses this information to correct the original GPS signals to eliminate or reduce the errors of satellite signals, such as ionospheric delay, tropospheric delay, etc.; after PTK processing, the inspection vehicle can obtain a higher positioning accuracy than using GPS alone. The inspection vehicle fuses the accurate position information obtained by the PTK positioning module with other sensor data to achieve more accurate berth positioning; finally, the inspection vehicle transmits the positioning information, license plate, and parking space information to the operation and management platform through a mobile communication module for final processing and analysis. The system includes: an image acquisition module, which is designed to obtain a berth image through a camera, wherein the berth image includes the license plate number of the vehicle and the berth number; a vehicle positioning module, which is designed to obtain the positioning information of the video inspection vehicle through a GPS positioning device; a data processing module, which is designed to preprocess the berth image and analyze the image data according to an image algorithm to obtain the license plate information and berth information, and determine the berth corresponding to the captured image based on the time of the video frame, the time corresponding to the license plate capture, and the position of the berth number in the image, wherein the image algorithm includes an OCR character recognition algorithm; a mobile communication module, which is designed to transmit the license plate and parking space information to the operation and management platform through mobile communication technology.

[0026] The PTK-based inspection vehicle berth positioning system provided by the present invention specifically includes: The license plate number of the vehicle is obtained by real-time shooting and recognition through an intelligent license plate recognition camera installed on the vehicle inspection vehicle; the berth number is obtained by real-time shooting and recognition through a special camera installed on the vehicle inspection vehicle for collecting a clear berth number video stream; the intelligent license plate recognition camera and the special camera are installed at the front end of the video inspection vehicle according to the relative positions of the license plate of the vehicle and the berth number.

[0027] After receiving the video stream collected by the berth number acquisition camera, the data processing module obtains the berth number data through the built-in OCR character recognition algorithm in the data processing module. It preprocesses the frame in the video stream, segments the berth number characters in the frame, and recognizes the content of the berth number characters, so as to obtain the berth number data. The position of the berth number acquisition camera is as Figure 2 shown.

[0028] The preprocessing of the frame in the video stream by the data processing module specifically includes: marking the text area, determining the position of the text in each document by marking the text area of the collected sample data, so as to construct a model to learn the spatial structure of the text and improve the recognition accuracy of the text; extracting the text content, extracting the text content in the marked text area and establishing the label data corresponding to the text content for supervised learning; data cleaning and preprocessing, cleaning and preprocessing the extracted text content, including removing noise, adjusting the size and direction, etc., to ensure the quality of the text data.

[0029] The OCR character recognition algorithm supports multiple models, including traditional template matching algorithms, support vector machines, Bayesian classification algorithms, and neural network models based on deep learning, etc. It can select a suitable model for training according to specific application scenarios and requirements; after the model training is completed, the data processing module improves the accuracy of the OCR character recognition algorithm through dataset optimization, feature selection, and model structure optimization. Among them, dataset optimization is specifically to use a more accurate and comprehensive dataset for training, and optimize the dataset by increasing the sample size, deleting duplicate samples, increasing randomness, etc.; feature selection is specifically to select the most important features for the recognition task and reduce unnecessary features, so as to speed up the response speed of the OCR system. The selection of relevant features is based on feature evaluation methods, such as information gain, correlation coefficient, etc.

[0030] The spraying of the berth number is based on the number spraying standard dedicated to the custom OCR system. The number spraying standard involves the font, font size, and color of the berth number, so as to improve the applicability of the berth number OCR algorithm. The berth number spraying standard is as Figure 3 shown, and the effect of the OCR character recognition algorithm is as Figure 4 shown.

[0031] The data processing module determines whether the berth information of this berth is correct by comprehensively comparing the berth information of the previous berth, the positioning information of the video inspection vehicle with the berth information of this berth and the positioning information of the video inspection vehicle. For example, if the berth number of the previous berth is 00050 and the physical distance between the positioning of the video inspection vehicle at the previous berth and the positioning of the video inspection vehicle at this berth is only 6 meters, but the calculated berth number of this berth is 00031, then it is judged that there may be a situation where the berth number of this berth cannot be correctly identified. At this time, comprehensive verification can be carried out according to the berth information and the positioning information of the video inspection vehicle of the next berth. If the video inspection vehicle only moves about 6 meters and then identifies the berth number of the next berth as 00052, then it can be basically determined that the actual berth number of this berth should be 00051. The reason for the failure to correctly identify the berth number of this berth may be that the berth number is blocked, contaminated or damaged. Similar abnormal situations can be submitted to the operation management platform for subsequent processing, assisting the image algorithm to judge the license plate information and berth information from multiple angles, improving the recognition and positioning accuracy, and improving the operation management efficiency.

[0032] The mobile communication module establishes a parking space geographic information database based on the geographic information scheme, which displays the location distribution map of parking berths in the form of a map, including: matching the data recognized by the OCR system with the geographic information in the geographic information scheme database, and the geographic information scheme database should contain information such as the geographic coordinates, attribution area, type, etc. of the berth; combining the real-time collected data with the geographic information scheme, and performing spatial analysis and processing, and using the geographic information scheme to monitor and manage the vehicle position and berth status in real time; displaying the berth status, vehicle position and relevant statistical information through the map view. The effect of the geographic information scheme is as Figure 5 shown.

[0033] Embodiment 2: On the basis of the foregoing embodiment, a deep learning scheme is added to improve the recognition accuracy and speed of the OCR character recognition algorithm. It uses a convolutional neural network as the basic model, and continuously improves the recognition ability of the OCR character recognition algorithm by continuously increasing the training data; Among them, it is necessary to preprocess the data in advance to normalize the image data to a unified range, such as or , ; Among them, represents the original image pixel value; represents the minimum value of the image pixel value, usually 0; represents the maximum value of the image pixel value, usually 255; represents the normalized pixel value, and the range is between 0 and 1.

[0034] Building a convolutional neural network includes: Applying multiple convolutional kernels in the convolutional layer to extract image features:

[0035] Among them, represents the output of the convolutional kernel ; represents the weight of the convolutional kernel; represents the pixel value of the input image; represents the bias; represents the total number of convolutional kernels; Using as the non-linear activation function:

[0036] Reducing the dimension of the feature map by using max pooling:

[0037] Among them, represents an element on the feature map output by the pooling layer; represents the element on the input feature map, and define the position within the pooling window; Flattening the feature map and connecting it to the output layer:

[0038] Among them, represents the output of the th neuron; represents the weight; represents the input feature; represents the bias term of the fully connected layer; represents the length of the input feature vector.

[0039] Loss function for multi-classification problems:

[0040] Among them, represents the value of the loss function, measuring the difference between the model prediction and the true label; represents the encoding of the true label; represents the probability distribution predicted by the model; represents the total number of classes.

[0041] Calculating the gradient of the loss function with respect to the model parameters:

[0042] Among them, Denote the gradient of the loss function with respect to the weights ; Denote the gradient of the loss function with respect to the output of the fully connected layer ; Denote the gradient of the activation function with respect to its input ; Denote the gradient of the input of the fully connected layer with respect to the weights ; And update the model parameters using gradient descent:

[0043] where, denotes the learning rate; denotes the weights in the next iteration; denotes the weights in the current iteration.

[0044] Prevent overfitting through L2 regularization:

[0045] where, denotes the regularization coefficient, which controls the strength of the regularization term.

[0046] For example, there is a dataset containing 1000 pictures of berth numbers, each picture is pixels in size, the pictures are grayscale, and the pictures have been preprocessed and normalized to the range, and there are 10 possible digits for the berth number, from 0 to 9; Suppose there is a picture of a berth number , whose pixel values range from 0 to 255, and normalize it to the range:

[0047] Build a convolutional neural network model, including: Convolutional layer: 32 convolutional kernels, with a size of , and use the ReLU activation function; Pooling layer: Max pooling, with a window size of ; Fully connected layer: 128 neurons, also using the ReLU activation function; Output layer: 10 neurons, corresponding to 10 digital categories; Use the cross-entropy loss function. Suppose the true label is , and the model prediction is :

[0048] Assume the learning rate is 0.01. After one forward pass and loss calculation, compute the gradients and update the weights:

[0049]

[0050] Adopt L2 regularization with a regularization coefficient of 0.001:

[0051]

[0052] The effect of the deep learning solution is shown in Table 1 as follows: Table 1. Summary of the effects of the deep learning solution Training cycle Training loss Training accuracy Validation loss Validation accuracy 1 2.30 0.20 2.35 0.18 2 1.50 0.45 1.60 0.40 3 1.10 0.60 1.25 0.55 4 0.90 0.70 1.10 0.60 5 0.75 0.75 0.95 0.65 6 0.65 0.80 0.85 0.70 7 0.55 0.85 0.75 0.75 As shown in the table, as the training progresses, the training loss gradually decreases and the training accuracy gradually increases, indicating that the performance of the model on the training data is improving; at the same time, the validation loss and validation accuracy also show that the performance of the model on unseen data is improving, but the improvement amplitude is not as large as that on the training set.

[0053] Those skilled in the art should understand that the embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable non-transitory storage media containing computer-usable program code.

[0054] The present invention can provide computer program instructions to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for implementing the system.

[0055] These computer program instructions can also be stored in a computer-readable memory capable of guiding the computer or other programmable data processing devices to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including an instruction device, and the instruction device implements the functions of the system.

[0056] These computer program instructions can also be loaded onto the computer or other programmable data processing devices, such that a series of operation steps are executed on the computer or other programmable devices to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable devices provide steps for implementing the functions of the system.

Claims

1. Inspection vehicle parking space positioning system, characterized by: include: The image acquisition module is designed to acquire a berth image through a camera, wherein the berth image Include the vehicle's license plate number and parking space number; The vehicle positioning module is designed to obtain the positioning information of the video inspection vehicle through the GPS positioning device; The data processing module is designed to pre-process the parking space image and analyze the image data according to the image algorithm to obtain the license plate information and the parking space information, and determine the parking space corresponding to the captured image according to the time of the picture frame, the time corresponding to the license plate shooting and the position of the parking space number in the image; wherein the image algorithm includes an OCR text recognition algorithm; The mobile communication module is designed to transmit license plate and parking space information to the operation management platform through mobile communication technology; The system adds a deep learning solution to improve the recognition accuracy and speed of the OCR text recognition algorithm. It uses a convolutional neural network as the basic model and continuously improves the recognition ability of the OCR text recognition algorithm by increasing the amount of training data.

2. The parking space positioning system for patrol vehicles according to claim 1, characterized in that: The vehicle's license plate number is captured and identified in real time by an intelligent license plate recognition camera installed on the vehicle inspection vehicle; the berth number is captured and identified in real time by a special camera installed on the vehicle inspection vehicle for collecting a clear berth number video stream; the intelligent license plate recognition camera and the special camera are installed at the front end of the video inspection vehicle according to the relative positions of the vehicle's license plate and the berth number.

3. The parking space positioning system for patrol vehicles according to claim 1 or 2, characterized in that: After receiving the video stream captured by the berth number acquisition camera, the data processing module obtains the berth number data through the OCR text recognition algorithm built into the data processing module. It obtains the berth number data by pre-processing the picture frames in the video stream, segmenting the berth number characters in the picture frames, and identifying the content of the berth number characters.

4. The parking space positioning system for patrol vehicles according to claim 3 is characterized in that: The data processing module preprocesses the picture frames in the video stream, specifically including: marking text areas, determining the location of text in each document by marking text areas in the collected sample data; extracting text content, extracting text content in the marked text area and establishing label data corresponding to the text content; data cleaning and preprocessing, cleaning and preprocessing the extracted text content, including removing noise, adjusting size and direction.

5. The parking space positioning system for patrol vehicles according to claim 3 is characterized in that: The OCR text recognition algorithm supports multiple models, including traditional template matching algorithms, support vector machines, Bayesian classification algorithms, and neural network models based on deep learning. It can select appropriate models for training according to specific application scenarios and needs. After the model training is completed, the data processing module improves the accuracy of the OCR text recognition algorithm through data set optimization, feature selection, and model structure optimization.

6. The parking space positioning system for patrol vehicles according to claim 5, characterized in that: The data processing module determines whether the berth information of the current berth is correct by comprehensively comparing the berth information of the previous berth and the positioning information of the video inspection vehicle with the berth information of the current berth and the positioning information of the video inspection vehicle.

7. The parking space positioning system for patrol vehicles according to claim 1, characterized in that: The mobile communication module establishes a parking space geographic information database based on a geographic information solution, which displays the location distribution diagram of parking spaces in the form of a map.

8. The parking space positioning system for patrol vehicles according to claim 8 is characterized in that Specifically include: Match the data recognized by the OCR system with the geographic information in the geographic information solution database, which should contain information such as the geographic coordinates, region, and type of the berth; Combine the real-time collected data with geographic information solutions, perform spatial analysis and processing, and use geographic information solutions to monitor and manage vehicle location and berth status in real time; display berth status, vehicle location and related statistical information through map views.

9. A method for locating a parking space for an inspection vehicle, characterized in that: The method is implemented based on the system according to any one of claims 1 to 8: The method comprises: The berth image is obtained through the camera installed on the video inspection vehicle; Get the positioning information of the video inspection vehicle through GPS positioning equipment; Preprocessing the berth image and analyzing the image data according to the image algorithm; Output license plate and parking space information through mobile communication technology.

10. A computer-readable storage medium, wherein a computer program is stored in the computer-readable storage medium, characterized in that: The computer program, when executed, implements the inspection vehicle parking space positioning method described in claim 9.

Images