Intelligent parking lot license plate recognition and parking space management system based on deep learning

Abstract

The invention relates to the technical field of intelligent parking lot vehicles, and discloses an intelligent parking lot license plate recognition and parking space management system based on deep learning, which comprises a comprehensive management platform, a license plate recognition module and a parking space management module, and is characterized in that the license plate recognition module is used for acquiring a license plate information set and analyzing a license plate according to the license plate information set; the parking space management module is used for obtaining a parking lot unoccupied parking space data set, carrying out parking space analysis according to parking space data, forming an optimal parking space, marking the optimal parking space as a vehicle destination, and finally generating an optimal parking path and sending the optimal parking path to the comprehensive management platform. The comprehensive management platform carries out management and distribution according to the received information, the license plate can be recognized more accurately, the situation that recognition cannot be carried out or recognition is mistakenly caused by the fact that the whole image is too dark is avoided, meanwhile, a proper parking position can be rapidly located, parking search time is greatly saved, and parking efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent parking lots, and in particular to an intelligent parking lot license plate recognition and parking space management system based on deep learning. Background Art

[0002] In modern society, with the continuous increase in the number of cars, the problem of parking difficulties has become increasingly prominent, becoming a major problem for urban development and people's lives. The emergence of smart parking lots, like a morning star, has brought a new dawn to solving this problem. With the help of cutting-edge technology, it deeply integrates vehicle management, information interaction, automatic control and other functions, and comprehensively reshapes all aspects of parking, making parking more convenient, efficient and intelligent, greatly improving people's travel experience.

[0003] In traditional parking lots, there are many manual management links, including card issuance, charging, guidance, etc., which are not only prone to human errors, but also have low work efficiency. Smart parking lots have realized intelligent and information-based vehicle management through automated equipment and systems, greatly reducing the need for manual intervention. Vehicles entering and exiting do not require manual operation. The system automatically identifies, charges, and releases vehicles, greatly improving the turnover rate of vehicles and increasing the throughput of parking lots. At the same time, the intelligent management system's reasonable allocation of parking space resources and effective monitoring of equipment also help reduce operating costs and improve the overall economic benefits of parking lots.

[0004] Although smart parking lots provide great convenience to people in daily life, there are also some problems in daily use. For example, at night or in poorly lit underground parking lots, the light reflected by the license plate is weak, and the camera cannot clearly capture the license plate image. At the same time, in strong sunlight, reflections may appear on the surface of the license plate, making it difficult for the camera to focus and accurately identify the license plate content. When part of the license plate is blocked by shadows and other parts are in bright places, it will cause uneven light, making it difficult for the recognition system to correctly analyze the license plate information. Summary of the invention

[0005] 1. Technical issues to be resolved

[0006] In view of the shortcomings of the prior art, the present invention provides an intelligent parking lot license plate recognition and parking space management system based on deep learning, which can more accurately identify license plates, avoid the unrecognizable or misrecognized situations caused by light problems or shadow problems, and at the same time, can quickly locate the appropriate parking position, greatly saving parking search time and improving parking efficiency.

[0007] (II) Technical solution

[0008] To achieve the above object, the present invention provides the following technical solutions: An intelligent parking lot license plate recognition and parking space management system based on deep learning, including an integrated management platform, a license plate recognition module, and a parking space management module. The license plate recognition module and the parking space management module are respectively connected to the integrated management platform through a network;

[0009] The license plate recognition module includes a license plate acquisition unit, a license plate analysis unit, and a license plate recognition unit. The license plate acquisition unit is used to acquire a license plate information set and send the license plate information set to the license plate analysis unit. The license plate analysis unit analyzes the license plate according to the license plate information set and generates a unique license plate number and sends it to the license plate recognition unit. The license plate recognition unit is internally provided with a license plate recognition system. The license plate recognition system recognizes the license plate according to the unique license plate number and sends the license plate recognition information to the integrated management platform;

[0010] The parking space management module includes a parking space acquisition unit, a parking space analysis unit, and a parking path planning unit. The parking space acquisition unit is used to acquire a dataset of available parking spaces in the parking lot and send the parking space data to the parking space analysis unit. The parking space analysis unit analyzes the parking space according to the parking space data and generates the best parking space and sends it to the parking path planning unit. The parking path planning unit is internally provided with a map system. The map system marks the best parking space as the vehicle destination and generates the best parking path and sends it to the integrated management platform;

[0011] The integrated management platform performs management and allocation according to the received information.

[0012] Preferably, the license plate information set is a plurality of license plate image data obtained by connecting a plurality of monitoring cameras in the intelligent parking lot through the license plate acquisition unit for shooting. The expression of the license plate information set is: , where represents the first data information in the license plate information set, that is, the first license plate image data captured by the first monitoring camera, represents the th data information in the license plate information set, that is, the th license plate image data captured by the th monitoring camera, represents the total number of data in the license plate information set, that is, the total number of monitoring cameras and the total number of license plate image data.

[0013] Preferably, the specific steps for the license plate analysis unit to analyze the license plate according to the license plate information set and generate a unique license plate number and send it to the license plate recognition unit are:

[0014] (1), separately cut each license plate image data in the license plate information set into pieces Mark the cut license plate image data with pictures of the appropriate size;

[0015] (2) Calculate the clarity of the picture information for each 、 、 、 、 photo;

[0016] (3) Sort the calculation results of the clarity of the picture information for each picture, and mark the picture with the highest clarity value of the picture information as the final picture of the nd picture. Then merge the to final pictures in sequence to form a unique license plate number.

[0017] Preferably, the marking of the cut license plate image data is as follows: 、 、 、 、 , where represents the first picture of the cut license plate image data, represents the th picture of the cut license plate image data, and 、 、 、 、 each have pictures.

[0018] Preferably, the calculation formula for the clarity of the picture information is;

[0019]

[0020] In the calculation formula, represents the clarity of the picture information of the th picture, which is the th picture in the th picture, represents the edge sharpness of the th picture in the th picture, and represents the average edge sharpness of the

[0021] Preferably, the formula for calculating the average edge sharpness is as follows;

[0022]

[0023] In the calculation formula, represents from to end, every one picture, to one the total value of the edge sharpness of the picture, represents every one the number of pictures.

[0024] Preferably, the dataset of available parking spaces in the parking lot is the data of available parking spaces obtained by connecting multiple monitoring cameras in the intelligent parking lot through the parking space acquisition unit. The expression of the dataset of available parking spaces in the parking lot is: where, represents the first empty space, represents the last empty space, represents the total quantity in the dataset of available parking spaces in the parking lot, that is, the total available parking spaces.

[0025] Preferably, the specific method of parking space analysis is as follows:

[0026] (1), Obtain the driving distance data between the current vehicle and the empty space and mark it as ;

[0027] (2), Calculate the parking influence number of each empty space according to the distance data and mark it as ;

[0028] (3), Mark the data with the lowest parking influence number as the best parking space.

[0029] Preferably, the calculation formula of the parking influence number is:

[0030]

[0031] In the calculation formula, represents the parking influence number, represents the fuel consumption, represents the number of oncoming vehicles, represents the number of vehicles with a speed lower than that of the current vehicle in the same direction.

[0032] Preferably, the integrated management platform sends the license plate recognition information to the intelligent parking lot management platform and sends the best parking path to the vehicle display platform.

[0033] Compared with the prior art, the present invention provides an intelligent parking lot license plate recognition and parking space management system based on deep learning, which has the following beneficial effects:

[0034] 1. By splitting the license plate image data in the license plate information set into multiple pictures, calculating the clarity of each picture, extracting the picture with the highest clarity as the final picture, and finally combining multiple final pictures into a unique license plate number, the present invention can focus on relatively clear areas, thereby more accurately recognizing the license plate, avoiding the situation of unrecognizable or misrecognized due to the overall image being too dark, and at the same time can obtain relatively clear license plate information, reducing recognition errors caused by uneven light, reducing misrecognition caused by environmental interference factors, and making the recognition result more accurate and reliable.

[0035] 2. By combining fuel consumption, the number of oncoming vehicles, and the number of vehicles with a speed lower than the vehicle's own speed in the same direction to calculate the parking impact number, and then marking the data with the lowest parking impact number as the best parking space, the present invention can quickly locate a suitable parking position, greatly saving the parking search time and improving the parking efficiency. It not only saves fuel costs for vehicle owners but also reduces exhaust emissions generated by fuel consumption, which is beneficial to environmental protection. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 It is a schematic diagram of the system of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0037] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0038] Please refer to Figure 1 , an intelligent parking lot license plate recognition and parking space management system based on deep learning, including a comprehensive management platform, a license plate recognition module, and a parking space management module. The license plate recognition module and the parking space management module are respectively connected to the comprehensive management platform through a network;

[0039] The license plate recognition module includes a license plate acquisition unit, a license plate analysis unit, and a license plate recognition unit. The license plate acquisition unit is used to acquire a license plate information set and send the license plate information set to the license plate analysis unit;

[0040] The license plate information set is a plurality of license plate image data obtained by connecting multiple monitoring cameras in the intelligent parking lot through the license plate acquisition unit. The expression of the license plate information set is: , where Represents the first data information in the license plate information set, that is, the first license plate image data captured by the first surveillance camera, Represents the th data information in the license plate information set, that is, the th license plate image data captured by the th surveillance camera, Indicates the total number of data in the license plate information set, that is, the total number of surveillance cameras and the total number of license plate image data;

[0041] The license plate information set is obtained by shooting with multiple surveillance cameras. Considering the shooting angles and environments of individual surveillance cameras, it is easy to have unclear license plate shooting during the license plate shooting process. By combining multiple license plate image data, the license plate is analyzed to improve the accuracy of license plate analysis and recognition;

[0042] The license plate analysis unit analyzes the license plate according to the license plate information set and generates a unique license plate number and sends it to the license plate recognition unit;

[0043] The specific steps for the license plate analysis unit to analyze the license plate according to the license plate information set and generate a unique license plate number and send it to the license plate recognition unit are as follows:

[0044] (1) Cut each license plate image data in the license plate information set separately into -sized pictures, and mark the cut license plate image data as: 、 、 、 、 、 , where represents the first picture after cutting the license plate image data, represents the th picture after cutting the license plate image data, represents the corresponding picture data after cutting all license plate image data in the license plate information set, that is, each 、 、 、 、 all have pictures;

[0045] (2) Calculate the clarity of the picture information of each 、 、 、 、 photo;

[0046]

[0047] In the calculation formula, represents the th in the picture, the clarity of the picture information of the th picture, represents the th in the picture, the edge sharpness of the th picture, represents the average edge sharpness of the th picture. The calculation formula for the average edge sharpness is as follows;

[0048]

[0049] In the calculation formula, represents from to ending, every th picture, to the total value of the edge sharpness of the th picture, represents the number of every th picture;

[0050] (3), Sort the calculation results of the picture information clarity of every th picture. Mark the picture with the highest picture information clarity value as the final picture of the th picture. Combine the to final pictures in sequence to form a unique license plate number;

[0051] Divide the license plate into multiple pictures, then calculate the clarity of each picture, extract the picture with the highest clarity as the final picture, and finally combine multiple final pictures into a unique license plate number, which can focus on relatively clearer areas, thus more accurately identifying the license plate, avoiding the situation of unrecognizable or misrecognized due to the overall image being too dark, and at the same time being able to obtain relatively clear license plate information, reducing recognition errors caused by uneven light, and reducing misrecognition caused by environmental interference factors, making the recognition result more accurate and reliable;

[0052] The license plate recognition unit is equipped with a license plate recognition system. The license plate recognition system recognizes the license plate according to the unique license plate number and sends the license plate recognition information to the integrated management platform;

[0053] The parking space management module includes a parking space acquisition unit, a parking space analysis unit, and a parking path planning unit. The parking space acquisition unit is used to acquire the dataset of available parking spaces in the parking lot and send the parking space data to the parking space analysis unit;

[0054] The dataset of available parking spaces in the parking lot is the data of available parking spaces acquired by connecting multiple monitoring cameras in the intelligent parking lot through the parking space acquisition unit. The expression of the dataset of available parking spaces in the parking lot is: , where represents the first empty space, represents the last empty space, represents the total number in the dataset of available parking spaces in the parking lot, that is, the total number of available parking spaces;

[0055] The parking space analysis unit analyzes the parking spaces based on the parking space data and generates the best parking space and sends it to the parking path planning unit;

[0056] The specific method of parking space analysis is:

[0057] (1) Obtain the driving distance data between the current vehicle and the empty space and mark it as ;

[0058] (2) Calculate the parking influence number of each empty space according to the distance data and mark it as ;

[0059]

[0060] In the calculation formula, represents the parking influence number, represents the fuel consumption, represents the number of oncoming vehicles, represents the number of vehicles with a speed lower than that of the vehicle in the same direction;

[0061] (3) Mark the data with the lowest parking influence number as the best parking space;

[0062] The parking path planning unit is internally equipped with a map system. The map system marks the best parking space as the destination of the vehicle and generates the best parking path and sends it to the integrated management platform;

[0063] By combining the fuel consumption, the number of oncoming vehicles, and the number of vehicles with a speed lower than that of the vehicle in the same direction to calculate the parking influence number, and then marking the data with the lowest parking influence number as the best parking space, it can quickly locate the appropriate parking position, greatly save the parking search time, improve the parking efficiency, not only save the fuel cost for the vehicle owner, but also reduce the exhaust emissions generated by fuel consumption, which is beneficial to environmental protection;

[0064] The integrated management platform sends the license plate recognition information to the intelligent parking lot management platform and sends the optimal parking path to the vehicle display platform.

[0065] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent parking lot license plate recognition and parking space management system based on deep learning, characterized by: It includes a comprehensive management platform, a license plate recognition module and a parking space management module, wherein the license plate recognition module and the parking space management module are respectively connected to the comprehensive management platform through a network; The license plate recognition module includes a license plate acquisition unit, a license plate analysis unit and a license plate recognition unit. The license plate acquisition unit is used to acquire a license plate information set and send the license plate information set to the license plate analysis unit. The license plate analysis unit analyzes the license plate according to the license plate information set, generates a unique license plate number and sends it to the license plate recognition unit. The license plate recognition unit is equipped with a license plate recognition system. The license plate recognition system recognizes the license plate according to the unique license plate number and sends the license plate recognition information to the integrated management platform. The parking space management module includes a parking space acquisition unit, a parking space analysis unit and a parking path planning unit. The parking space acquisition unit is used to acquire a data set of available parking spaces in the parking lot and send the parking space data to the parking space analysis unit. The parking space analysis unit performs parking space analysis based on the parking space data and generates an optimal parking space and sends it to the parking path planning unit. The parking path planning unit is internally provided with a map system. The map system marks the optimal parking space as the vehicle destination and generates an optimal parking path and sends it to the integrated management platform. The integrated management platform performs management allocation according to the received information.

2. According to claim 1, a deep learning-based intelligent parking lot license plate recognition and parking space management system is characterized by: The license plate information set is a plurality of license plate image data captured by connecting a plurality of surveillance cameras in a smart parking lot through a license plate acquisition unit. The expression of the license plate information set is: ,in, Represents the first data information in the license plate information set, that is, the first license plate image data taken by the first surveillance camera. Represents the first data information, that is, through the The surveillance camera captured License plate image data, Represents the total data quantity of the license plate information set, that is, the total number of surveillance cameras and the total number of license plate image data.

3. According to claim 2, the intelligent parking lot license plate recognition and parking space management system based on deep learning is characterized by: The specific steps of the license plate analysis unit analyzing the license plate according to the license plate information set and generating a unique license plate number to send to the license plate recognition unit are: (1) Each license plate image data in the license plate information set is cut into indivual The image of the license plate after cutting is marked. (2) Calculate each , , , , The clarity of the picture information of the photo; (3) indivual The pictures are sorted by the calculation results of the picture information clarity, and the picture with the highest picture information clarity value is marked as indivual The final image of the picture will be arrive The final images are merged in sequence to form a unique license plate number.

4. According to claim 3, the intelligent parking lot license plate recognition and parking space management system based on deep learning is characterized by: The cut license plate image data is marked as follows: , , , , ,in, The first image representing the license plate image data to be cut. Represents the license plate image data for cutting pictures, Represents the data of the corresponding picture after cutting all the license plate image data in the license plate information set, that is, each , , , , Both pictures.

5. According to claim 4, a deep learning-based intelligent parking lot license plate recognition and parking space management system is characterized by: The calculation formula of the picture information clarity is: In the calculation formula, Representative indivual In the picture, The clarity of the image information of the image, Representative indivual In the picture, The edge sharpness of the image, Representative indivual The average edge sharpness of the image.

6. According to claim 5, the deep learning-based intelligent parking lot license plate recognition and parking space management system is characterized by: The average edge sharpness calculation formula is as follows: In the calculation formula, Representative from Start to End, every indivual Pictures, to indivual The total value of the edge sharpness of the image, Represents each indivual The number of images.

7. According to claim 6, a deep learning-based intelligent parking lot license plate recognition and parking space management system is characterized by: The parking lot vacant parking space dataset is the vacant parking space data acquired by connecting a plurality of surveillance cameras in the smart parking lot through the parking space acquisition unit. The expression of the parking lot vacant parking space dataset is: ,in, Represents the first vacancy, Represents the last empty position. Represents the total number of vacant parking spaces in the parking lot dataset, that is, the total vacant parking spaces.

8. The deep learning-based intelligent parking lot license plate recognition and parking space management system according to claim 7 is characterized by: The specific method of the parking space analysis is as follows: (1) Obtain the driving distance data between the current vehicle and the empty space, marked as ; (2) Calculate the parking impact number of each vacant space based on the distance data and mark it as ; (3) The number of parking impacts The lowest data is marked as the best parking spot.

9. The deep learning-based intelligent parking lot license plate recognition and parking space management system according to claim 8, characterized in that: The calculation formula of the parking impact number is: In the calculation formula, represents the parking impact number, Represents fuel consumption, Represents the number of oncoming vehicles. Represents the number of vehicles traveling in the same direction at a lower speed than the vehicle itself.

10. The deep learning-based intelligent parking lot license plate recognition and parking space management system according to claim 9, characterized in that: The integrated management platform sends the license plate recognition information to the intelligent parking lot management platform and sends the best parking path to the car display platform.

Images