Expressway closed service area traffic flow and passenger flow accurate monitoring and early warning service system and method

Abstract

The invention relates to the field of expressway intelligent service areas, and mainly relates to an expressway closed service area traffic flow and passenger flow accurate monitoring and early warning service system and method, and the system comprises a service area entrance and exit ramp sensing module, a fusion recognition module, a monitoring research and judgment module, an early warning information issuing module and a client service module. Video cameras and radar equipment are deployed at entrances and exits of a closed service area, accurate acquisition, identification, research and judgment, early warning, information release and client service of the number of vehicles entering and exiting the service area, license plates, vehicle types, license plate features and passengers in the vehicles are realized based on a deep learning algorithm, and graded and classified early warning service of the traffic flow and the people flow in the service area is realized. Comprehensive management and control of the service area are enhanced; the current situation information of parking spaces and passenger flow congestion degree of a front service area is provided for travelers, so that the travelers can make scientific decisions on whether to enter or not, safety and orderliness of vehicles and personnel in the service area are ensured, and the service quality is ensured.

Description

Technical Field

[0001] The present invention relates to the field of smart highway service areas, and in particular to a system and method for accurately monitoring and warning vehicle and passenger flows in closed highway service areas. Background Art

[0002] With the rapid expansion of the highway network, service areas, as important places on highways that provide services such as parking, refueling, dining and rest, are being used more and more frequently. Especially during holidays or peak traffic periods, the traffic and passenger flow in service areas increases sharply, leading to problems such as tight parking spaces and crowded people. This not only reduces the service level of service areas, but also increases traffic safety risks.

[0003] Traditional highway service area management mainly relies on manual inspections and simple camera monitoring, which makes it difficult to achieve real-time, efficient monitoring and management of vehicle and passenger flows. At the same time, faced with drastic fluctuations in vehicle and passenger flows, service area managers often find it difficult to respond and adjust in a timely manner, which can easily lead to congestion, accidents and other problems, affecting the operating efficiency of the service area and the quality of travel services.

[0004] With the advancement of information technology, especially the widespread application of technologies such as the Internet of Things and big data analysis, intelligent management of highway service areas has become possible. Through accurate monitoring and data analysis of vehicle and passenger flows in the service area, real-time monitoring and early warning of the use of parking spaces in the service area can be achieved, thereby improving the operational management capabilities of the service area. However, most monitoring systems on the market currently have a large number of installed equipment and are costly, making it difficult to meet the widespread application needs of highway service areas across the country. Therefore, it is of great practical significance to develop a system that requires less equipment installation, is cost-effective, and has real-time monitoring and early warning functions. Summary of the invention

[0005] The purpose of the present invention is to provide a system and method for accurately monitoring and warning the traffic and passenger flow in a closed service area of ​​a highway, so as to solve the above-mentioned problems existing in the prior art.

[0006] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0007] A precise monitoring and early warning service system for vehicle and passenger flow in closed service areas of expressways, including a service area entrance and exit ramp sensing module, a fusion recognition module, a vehicle and passenger flow monitoring and analysis module, an early warning information release module and a client service module;

[0008] The service area entrance and exit ramp sensing module is composed of an entrance ramp sensing unit, an exit ramp sensing unit and a network transmission unit; the entrance ramp sensing unit is arranged on the entrance ramp pole of the closed service area, including an imported high-definition intelligent camera, an imported millimeter-wave radar, an imported thermal imaging camera, an imported infrared gas strobe light, and an imported LED fill light; the exit ramp sensing unit is arranged on the exit ramp pole of the service area, including an exit high-definition intelligent camera, an exit millimeter-wave radar, an exit thermal imaging camera, an exit infrared gas strobe light, and an exit LED fill light; the network transmission unit includes a network switch, a server, a storage terminal server, and a transmission network; the network transmission unit is used to system The setting position of the entrance ramp pole can ensure that the import high-definition smart camera on it can take panoramic photos of the front and side of the vehicle entering the area, the license plate and the model; after the photos are taken, they are uploaded to the fusion recognition module for identification; the number of people in the vehicle entering the area is counted by the import thermal imaging camera; the setting position of the exit ramp pole can ensure that the exit high-definition smart camera on it can take panoramic photos of the front and side of the vehicle leaving the area, the license plate and the model; after the photos are taken, they are uploaded to the fusion recognition module for identification; the number of people in the vehicle leaving the area is counted by the exit thermal imaging camera; the license plate information is the unique identification of each vehicle, which is used to identify the corresponding vehicle;

[0009] The fusion recognition module includes a vehicle front feature recognition unit, a vehicle side feature recognition unit, a key vehicle track recognition unit and a vehicle interior personnel recognition unit;

[0010] The vehicle and passenger flow monitoring and judgment module includes a vehicle flow early warning monitoring and judgment unit, a passenger flow early warning monitoring and judgment unit, and a key vehicle electronic fence unit;

[0011] The warning information release module is used to release warning information; the warning information is released in three ways: ① setting up an electronic display screen and broadcasting equipment 200-500 meters upstream of the service area entrance to release the warning information; ② releasing the warning information through the variable information board on the ETC gantry upstream of the service area; ③ releasing the warning information to the client service module;

[0012] The client service module is installed on the relevant user's mobile phone and / or car computer, and is used to determine whether to actively pop up the warning information according to the current location of the machine, and passively display the warning information and basic information of the service area according to user instructions.

[0013] Preferably, the specific process of determining whether to actively pop up the warning information based on the current location of the local machine is that after the client service module receives the warning information issued by the warning information issuing module, it first determines whether the current navigation route of the local machine passes through the service area, and then determines whether the local machine is within the coverage of the warning information. If both judgment results are yes, the warning information is automatically popped up through the local display device.

[0014] Preferably, the determining whether the local machine is within the coverage of the warning information is specifically determining whether the current location of the local machine is within a certain distance upstream of the service area.

[0015] A method for accurately monitoring and warning vehicle and passenger flow in a closed service area of ​​a highway is implemented based on the system and specifically includes the following steps:

[0016] S1, collecting the traffic information in and out of the closed service area in real time, and calculating the total number of existing vehicles in the service area through a closed service area real-time traffic status monitoring algorithm;

[0017] The step S1 specifically comprises: calculating the real-time traffic state of the closed service area through the real-time data of the entrance ramp sensing unit and the exit ramp sensing unit; setting the difference in the total number of vehicles entering and leaving the service area measured by the entrance and exit ramp sensing units to the total number of existing vehicles in the service area, and the device numbers of the entrance and exit ramp sensing units of the service area are respectively: the number of the i-th import vehicle, denoted as Cin i , i = 1-N, the number of vehicles at the jth exit, denoted as Cout j , j = 1-M, then the total number of existing vehicles in the service area is expressed by the formula:

[0018] S2, identifying and acquiring the type of existing vehicles in the service area based on the CNN-based vehicle side feature recognition algorithm according to the data obtained by the entrance and exit ramp sensing unit in step S1;

[0019] The CNN-based vehicle side feature recognition algorithm of step S2 specifically includes: using the convolutional neural network CNN in deep learning technology as the core model, and reducing the complexity by designing the FlowNet structure; the first is FlowNetHalf, which reduces the number of convolution kernels in each layer of FlowNet to half, and reduces the complexity to 1 / 4. The second is FlowNetInception, which uses the Inception network structure to reduce the complexity to 1 / 8, and automatically learns the method of automatically extracting and identifying the side features of the vehicle in the image through fast image processing, multi-layer convolution and pooling operations, thereby improving the accuracy and efficiency of recognition;

[0020] S3, using the closed parking lot real-time passenger flow status monitoring algorithm to calculate the maximum number of existing passenger flows;

[0021] The real-time passenger flow status monitoring algorithm of step S3 specifically includes: calculating the number of people in the existing vehicles in the service area through the data measured by the entrance ramp sensing unit and the exit ramp sensing unit, which is the real-time passenger flow number in the service area; binding according to the license plate information as a unique identifier, when the vehicle is monitored to leave, the number of passengers corresponding to the vehicle is also synchronously reduced, and the maximum number of existing passengers in the service area at time t can be obtained; then the real-time passenger flow status monitoring algorithm is expressed by the formula:

[0022]

[0023] Passengers t , represents the current number of passengers in the service area at time t;

[0024] It represents the jth vehicle of the sensing unit of the import ramp i at time t;

[0025] represents the number of passengers in the jth vehicle of the entrance / exit ramp sensing unit at time t;

[0026] It represents the jth vehicle of the ramp sensing unit at exit i at time t;

[0027] S4, counting the total number of parking spaces of various types in the service area, and then obtaining the total number of all parking spaces;

[0028] The step S4 specifically includes: the total number of parking spaces in the service area is recorded as Park max , the total number of parking spaces of various types in the service area is counted. The first type is small passenger car parking spaces, denoted as P Car The second category includes parking spaces for large buses and ordinary trucks, denoted by P Bus The third category is parking spaces for large trucks and hazardous chemicals transport vehicles, denoted by P Van ; The fourth category is parking spaces for other types of vehicles, denoted by P other ; then the total number of parking spaces in the service area Park max The calculation formula is: Park max =Σ(P Van +P Car +P Bus +P other );

[0029] S5, counting the upper limit of passenger flow capacity of each area in the service area, and then obtaining the maximum upper limit of passenger flow capacity of the service area;

[0030] The step S5 specifically includes: the maximum passenger flow capacity of the service area is recorded as Passengers max , is the sum of the passenger flow upper limit capacity of the comprehensive business area, public toilets, and other passenger flow areas within the service area. The calculation formula is:

[0031] Passengers max =∑(Passengers Business +Passengers toilet +Passengers other )

[0032] Where: Passengers Business Indicates the passenger flow capacity limit of the comprehensive business area, Passengers toilet Indicates the upper limit of the passenger flow capacity of the public toilet, Passengers other Indicates the upper limit of passenger flow capacity in other passenger flow areas;

[0033] S6, calculating the current congestion situation of the service area through a service area early warning analysis algorithm based on the above data;

[0034] The service area early warning analysis algorithm of step S6 includes: a real-time status algorithm for parking spaces in the service area and a real-time status algorithm for passenger flow in the service area;

[0035] In the service area parking space real-time status algorithm, the service area parking space real-time status is divided into three states: 0 idle; 1 normal; 2 crowded, and the algorithm is as follows:

[0036] R Vehicle idle =Car now / Park max , R Vehicle idle <0.6, indicating idle, the real-time status is displayed as 0 idle; 0.6≤R Vehicle idle <1.0, indicating normal, the real-time status is 1 normal, R Vehicle idle ≥1.0, indicating congestion, showing the real-time status as 2 congestion;

[0037] In the service area passenger flow real-time state algorithm, the service area passenger flow real-time state is divided into three states: 0 idle; 1 normal; 2 crowded, and the algorithm is as follows:

[0038] R passenger idle =Passengers now / Passengers max , R Passengers idle <0.6, indicating idle, the real-time status is displayed as 0 idle; 0.6≤R Passengers idle<1.0, indicating normal, the real-time status is 1 normal, R Passengers idle ≥1.0, indicating congestion, showing the real-time status as 2 congestion;

[0039] S7, constructing a vehicle electronic fence for key vehicles, and monitoring the path trajectory and residence time of the key vehicles in the service area, and triggering an early warning and taking preset measures if relevant regulations are violated;

[0040] The step S7 specifically includes: constructing a vehicle electronic fence in the parking area for hazardous chemicals transport vehicles in the service area, and giving an early warning of the parking position and parking time of hazardous chemicals transport vehicles entering the service area; binding the hazardous chemicals transport vehicle with its license plate information, and tracking the complete path trajectory of the hazardous chemicals transport vehicle in the service area; and using the GPS Beidou positioning data of key operating vehicles for data fitting to realize electronic fence management in the service area. If the hazardous chemicals transport vehicle enters the service area but does not enter the electronic fence corresponding to the parking area for hazardous chemicals transport vehicles, an early warning is issued, or if the vehicle is in the electronic fence for more than half an hour, the service area staff is notified to conduct on-site inquiries, and if it exceeds 1 hour, the owner of the vehicle and the company to which the vehicle belongs are notified to issue an overtime early warning, thereby realizing the full process monitoring of hazardous chemicals transport vehicles from entry to departure, and actively pushing monitoring information to the driver, owner and car company of the vehicle, issuing an overtime early warning and a telephone warning, and ending the full process monitoring when leaving the service area;

[0041] S8, display the service area warning information board through the electronic display screen and variable information board set up upstream of the service area entrance, and indicate the warning information and basic information of the service area; and prompt relevant users through information push or broadcast.

[0042] Preferably, the step S2 specifically includes:

[0043] ① Data preprocessing: Preprocess the vehicle side images obtained and transmitted by the entrance and exit ramp perception unit, including image scaling, normalization, and enhancement operations to improve the robustness of the model;

[0044] ② Feature extraction: Use the pre-trained CNN model ResNet or VGG or a custom CNN model to extract the features of the vehicle side image; these features include the vehicle's outline, windows, and wheels;

[0045] ③ Feature classification: The extracted features are input into the fully connected layer or classifier to classify the vehicle types; the classifier uses a simple Softmax classifier to classify the vehicle types entering the service area into Among them: 1 small passenger car, 2 large passenger car, 3 ordinary truck, 4 large truck, 5 hazardous chemical transport vehicle, 6 livestock transport vehicle, 7 others;

[0046] ④ Identify and count the number of people in the car: The thermal imaging camera in the entrance and exit ramp sensing unit takes side photos of the passing vehicles, and uses the deep learning algorithm to identify the number of vehicle occupants and count them. And bind it with the license plate;

[0047] ⑤ Taking K cars entering the parking lot at time T as an example, the number of passengers on K cars is:

[0048] Preferably, the content published and displayed on the electronic display screen and variable information board should comply with the relevant provisions of the current "Highway Speed ​​Limit Sign Design Specifications" and "Highway LED Variable Information Signs".

[0049] The beneficial effects of the present invention are:

[0050] The implementation of the technical solution of the present invention can conduct real-time monitoring and early warning of the vacant status of parking spaces in closed service areas of highways across the country, and can provide early warning of passenger congestion in service areas. The present invention has the characteristics of less equipment installation, high cost performance, and strong real-time performance, and can quickly improve the service area operation management capabilities and travel service levels. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 is a schematic diagram of the layout of the entrance ramp sensing unit in the present invention;

[0052] Figure 2 It is a schematic diagram of the system architecture of the present invention. DETAILED DESCRIPTION

[0053] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation methods described herein are only used to explain the present invention and are not used to limit the present invention.

[0054] The present invention relates to the field of smart service areas on highways, and is mainly applied to the implementation method of the precise monitoring and early warning service system for vehicle and passenger flow in closed smart service areas on highways, including a service area entrance and exit ramp sensing module, a fusion recognition module, a monitoring and judgment module, an early warning information release module, and a client service module. By deploying video cameras and radar equipment at the entrance and exit of closed service areas, the number of vehicles, license plates, vehicle models, license plate features, and passengers in the vehicle are accurately collected, identified, judged, warned, and information released, as well as client services based on deep learning algorithms to achieve graded and classified early warning services for vehicle and passenger flow in service areas, and strengthen the comprehensive management and control of service areas; provide travelers with information on the current status of parking spaces and passenger congestion in the service area ahead, so that travelers can make scientific decisions on whether to enter, ensure the safety and order of vehicles and personnel in the service area, and ensure service quality.

[0055] In accordance with the requirements of the "Basic Functional Requirements for the Digital Construction of Highway Service Areas (Trial)", the digital construction of highway service areas across the country needs to gradually realize the real-time monitoring, automatic collection and external release of service area data information, and improve the service area operation and management capabilities and travel service levels.

[0056] The service area vehicle and passenger flow precise monitoring and early warning service system based on the present invention is an important part of the construction of digital service areas. It can support the construction of provincial service area management information platforms and highway service area terminal systems (hereinafter referred to as terminal systems), which helps to improve the operational efficiency of service areas and enhance safety management capabilities.

[0057] The closed service area vehicle flow and passenger flow precise monitoring and early warning service system of the highway of the present invention comprises a service area entrance and exit ramp sensing module, a fusion recognition module, a vehicle flow and passenger flow monitoring and judgment module, an early warning information release module and a client service module; the service area entrance and exit ramp sensing module is composed of an entrance ramp sensing unit, an exit ramp sensing unit and a network transmission unit; the entrance ramp sensing unit is arranged on the entrance ramp pole of the closed service area, and comprises an imported high-definition intelligent camera, an imported millimeter-wave radar, an imported thermal imaging camera, an imported infrared gas strobe light and an imported LED fill light; the exit ramp sensing unit is arranged on the exit ramp pole of the service area, and comprises an exit high-definition intelligent camera, an exit millimeter-wave radar, an exit thermal imaging camera, an exit infrared gas strobe light and an exit LED fill light; the network The transmission unit includes a network switch, a server, a storage terminal server, and a transmission network; the network transmission unit is used for transmitting system-related data; the setting position of the entrance ramp pole can ensure that the import high-definition smart camera thereon can take panoramic photos of the front and side of the vehicle entering the area, the license plate and the model; after the photos are taken, they are uploaded to the fusion recognition module for identification; the number of people in the vehicle entering the area is counted by the import thermal imaging camera; the setting position of the exit ramp pole can ensure that the exit high-definition smart camera thereon can take panoramic photos of the front and side of the vehicle leaving the area, the license plate and the model; after the photos are taken, they are uploaded to the fusion recognition module for identification; the number of people in the vehicle leaving the area is counted by the exit thermal imaging camera; the license plate information is the unique identification of each vehicle, which is used to identify the corresponding vehicle;

[0058] The fusion recognition module includes a vehicle front feature recognition unit, a vehicle side feature recognition unit, a key vehicle track recognition unit and a vehicle interior personnel recognition unit;

[0059] The vehicle and passenger flow monitoring and judgment module includes a vehicle flow early warning monitoring and judgment unit, a passenger flow early warning monitoring and judgment unit, and a key vehicle electronic fence unit;

[0060] The warning information release module is used to release warning information; the warning information is released in three ways: ① Setting an electronic display screen and broadcasting equipment 200-500 meters upstream of the service area entrance to release the warning information; ② Release the warning information through the variable information board on the ETC gantry upstream of the service area; ③ Release the warning information to the client service module; the client service module includes mobile navigation software, service area APP, WeChat applets and other mobile client service modules; mobile navigation software, service area APP, WeChat applets and other mobile client service modules are used to release warning information and basic information for travelers, especially drivers on the way.

[0061] The client service module is installed on the relevant user's mobile phone and / or car computer, and is used to determine whether to actively pop up the warning information according to the current location of the machine, and passively display the warning information and basic information of the service area according to user instructions.

[0062] The specific process of judging whether to actively pop up the warning information according to the current position of the machine is that after the client service module receives the warning information issued by the warning information issuing module, it first judges whether the current navigation route of the machine passes through the service area, and then judges whether the machine is within the scope of the warning information. If both judgment results are yes, the warning information is automatically popped up through the display device of the machine. The judgment of whether the machine is within the scope of the warning information is specifically to judge whether the current position of the machine is within a certain distance upstream of the service area.

[0063] The method for accurately monitoring and warning vehicle and passenger flow in a closed service area of ​​a highway of the present invention is implemented based on the above system and specifically includes the following steps:

[0064] S1, collecting the traffic information in and out of the closed service area in real time, and calculating the total number of existing vehicles in the service area through a closed service area real-time traffic status monitoring algorithm;

[0065] The step S1 specifically comprises: calculating the real-time traffic state of the closed service area through the real-time data of the entrance ramp sensing unit and the exit ramp sensing unit; setting the difference in the total number of vehicles entering and leaving the service area measured by the entrance and exit ramp sensing units to the total number of existing vehicles in the service area, and the device numbers of the entrance and exit ramp sensing units of the service area are respectively: the number of the i-th import vehicle, denoted as Cin i , i = 1-N, the number of vehicles at the jth exit, denoted as Cout j , j = 1-M, then the total number of existing vehicles in the service area is expressed by the formula:

[0066] S2, identifying and acquiring the type of existing vehicles in the service area based on the CNN-based vehicle side feature recognition algorithm according to the data obtained by the entrance and exit ramp sensing unit in step S1;

[0067] By deploying cameras, infrared gas flash lights, LED fill lights, storage terminal servers and other equipment on the front and sides of the entrance and exit ramps of the service area, deep learning algorithms are used to identify vehicle models, license plates, the number of passengers in the vehicle (the maximum number of passengers can be verified by the model), vehicle color and other information;

[0068] The CNN-based vehicle side feature recognition algorithm of step S2 specifically includes: using the convolutional neural network CNN in deep learning technology as the core model, and reducing the complexity by designing the FlowNet structure; the first is FlowNetHalf, which reduces the number of convolution kernels in each layer of FlowNet to half, and reduces the complexity to 1 / 4. The second is FlowNetInception, which uses the Inception network structure to reduce the complexity to 1 / 8, and automatically learns the method of automatically extracting and identifying the side features of the vehicle in the image through fast image processing, multi-layer convolution and pooling operations, thereby improving the accuracy and efficiency of recognition;

[0069] The step S2 specifically includes:

[0070] ① Data preprocessing: Preprocess the vehicle side images obtained and transmitted by the entrance and exit ramp perception unit, including image scaling, normalization, and enhancement operations to improve the robustness of the model;

[0071] ② Feature extraction: Use the pre-trained CNN model ResNet or VGG or a custom CNN model to extract the features of the vehicle side image; these features include the vehicle's outline, windows, and wheels;

[0072] ③ Feature classification: The extracted features are input into the fully connected layer or classifier to classify the vehicle types; the classifier uses a simple Softmax classifier to classify the vehicle types entering the service area into Among them: 1 small passenger car, 2 large passenger car, 3 ordinary truck, 4 large truck, 5 hazardous chemical transport vehicle, 6 livestock transport vehicle, 7 others;

[0073] ④ Identify and count the number of people in the car: The thermal imaging camera in the entrance and exit ramp sensing unit takes side photos of the passing vehicles, and uses the deep learning algorithm to identify the number of vehicle occupants and count them. And bind it with the license plate;

[0074] ⑤ Taking K cars entering the parking lot at time T as an example, the number of passengers on K cars is:

[0075] S3, using the closed parking lot real-time passenger flow status monitoring algorithm to calculate the maximum number of existing passenger flows;

[0076] The real-time passenger flow status monitoring algorithm of step S3 specifically includes: calculating the number of people in the existing vehicles in the service area through the data measured by the entrance ramp sensing unit and the exit ramp sensing unit, which is the real-time passenger flow number in the service area; binding according to the license plate information as the unique identifier, when the vehicle is monitored to leave, the number of passengers corresponding to the vehicle is also synchronously reduced, and the maximum number of existing passengers in the service area can be obtained; then the real-time passenger flow status monitoring algorithm is expressed by the formula:

[0077]

[0078] Passengers t , represents the current number of passengers in the service area at time t;

[0079] It represents the jth vehicle of the sensing unit of the import ramp i at time t;

[0080] represents the number of passengers in the jth vehicle of the entrance / exit ramp sensing unit at time t;

[0081] It represents the jth vehicle of the ramp sensing unit at exit i at time t;

[0082] S4, counting the total number of parking spaces of various types in the service area, and then obtaining the total number of all parking spaces;

[0083] The step S4 specifically includes: the total number of parking spaces in the service area is recorded as Park max , the total number of parking spaces of various types in the service area is counted. The first type is small passenger car parking spaces, denoted as P Car The second category includes parking spaces for large buses and ordinary trucks, denoted by P Bus The third category is parking spaces for large trucks and hazardous chemicals transport vehicles, denoted by P Van ; The fourth category is parking spaces for other types of vehicles, denoted by P other ; then the total number of parking spaces in the service area Park max The calculation formula is: Park max =∑(P Van +P Car +P Bus +P other );

[0084] S5, counting the upper limit of passenger flow capacity of each area in the service area, and then obtaining the maximum upper limit of passenger flow capacity of the service area;

[0085] The step S5 specifically includes: the maximum passenger flow capacity of the service area is recorded as Passengers max , is the sum of the passenger flow upper limit capacity of the comprehensive business area, public toilets, and other passenger flow areas within the service area. The calculation formula is:

[0086] Passengers max =Σ(Passengers Business +Passengers toilet +Passengers other )

[0087] Where: Passengers Business Indicates the passenger flow capacity limit of the comprehensive business area, Passengers toilet Indicates the upper limit of the passenger flow capacity of the public toilet, Passengers other Indicates the upper limit of passenger flow capacity in other passenger flow areas;

[0088] S6, calculating the current congestion situation of the service area through a service area early warning analysis algorithm based on the above data;

[0089] The service area early warning analysis algorithm of step S6 includes: a real-time status algorithm for parking spaces in the service area and a real-time status algorithm for passenger flow in the service area;

[0090] In the service area parking space real-time status algorithm, the service area parking space real-time status is divided into three states: 0 idle; 1 normal; 2 crowded, and the algorithm is as follows:

[0091] R Vehicle idle =Car now / Park max , R Vehicle idle <0.6, indicating idle, the real-time status is displayed as 0 idle; 0.6≤R Vehicle idle <1.0, indicating normal, the real-time status is 1 normal, R Vehicle idle ≥1.0, indicating congestion, showing the real-time status as 2 congestion;

[0092] In the service area passenger flow real-time state algorithm, the service area passenger flow real-time state is divided into three states: 0 idle; 1 normal; 2 crowded, and the algorithm is as follows:

[0093] R passenger idle =Passengers now / Passengers max , R Passengers idle <0.6, indicating idle, the real-time status is displayed as 0 idle; 0.6≤R Passengers i dle<1.0, indicating normal, the real-time status is 1 normal, R Passengers idle ≥1.0, indicating congestion, showing the real-time status as 2 congestion;

[0094] S7, constructing a vehicle electronic fence for key vehicles, and monitoring the path trajectory and residence time of the key vehicles in the service area, and triggering an early warning and taking preset measures if relevant regulations are violated;

[0095] The step S7 specifically includes: constructing a vehicle electronic fence in the parking area for hazardous chemicals transport vehicles in the service area, and giving an early warning of the parking position and parking time of hazardous chemicals transport vehicles entering the service area; binding the hazardous chemicals transport vehicle with its license plate information, and tracking the complete path trajectory of the hazardous chemicals transport vehicle in the service area; and using the GPS Beidou positioning data of key operating vehicles for data fitting to realize electronic fence management in the service area. If the hazardous chemicals transport vehicle enters the service area but does not enter the electronic fence corresponding to the parking area for hazardous chemicals transport vehicles, an early warning is issued, or if the vehicle is in the electronic fence for more than half an hour, the service area staff is notified to conduct on-site inquiries, and if it exceeds 1 hour, the owner of the vehicle and the company to which the vehicle belongs are notified to issue an overtime early warning, thereby realizing the full process monitoring of hazardous chemicals transport vehicles from entry to departure, and actively pushing monitoring information to the driver, owner and car company of the vehicle, issuing an overtime early warning and a telephone warning, and ending the full process monitoring when leaving the service area;

[0096] S8, display the service area warning information board through the electronic display screen and variable information board installed upstream of the service area entrance, and indicate the warning information and basic information of the service area; and prompt relevant users through information push or broadcast. The content published and displayed by the electronic display screen and variable information board shall comply with the relevant provisions of the current "Highway Speed ​​Limit Sign Design Specifications" and "Highway LED Variable Information Signs". The above basic information can also be provided to the public through roadside broadcasting equipment or through smart terminals such as mobile phone navigation software, so as to reasonably guide the public to choose service areas and improve user satisfaction.

[0097] By adopting the above technical solution disclosed in the present invention, the following beneficial effects are obtained:

[0098] The implementation of the technical solution of the present invention can conduct real-time monitoring and early warning of the vacant status of parking spaces in closed service areas of highways across the country, and can provide early warning of passenger congestion in service areas. The present invention has the characteristics of less equipment installation, high cost performance, and strong real-time performance, and can quickly improve the service area operation management capabilities and travel service levels.

[0099] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be considered as the scope of protection of the present invention.

Claims

1. A precise monitoring and early warning service system for vehicle and passenger flow in closed service areas of expressways, characterized in that: It includes the service area entrance and exit ramp perception module, fusion recognition module, vehicle flow and passenger flow monitoring and analysis module, early warning information release module and client service module; The service area entrance and exit ramp sensing module is composed of an entrance ramp sensing unit, an exit ramp sensing unit and a network transmission unit; the entrance ramp sensing unit is arranged on the entrance ramp pole of the closed service area, and includes an imported high-definition intelligent camera, an imported millimeter-wave radar, an imported thermal imaging camera, an imported infrared gas flash light, and an imported LED fill light; The exit ramp sensing unit is arranged on the exit ramp pole of the service area, and includes an exit high-definition smart camera, an exit millimeter-wave radar, an exit thermal imaging camera, an exit infrared gas flash light, and an exit LED fill light; the network transmission unit includes a network switch, a server, a storage terminal server, and a transmission network; the network transmission unit is used for the transmission of system-related data; the setting position of the entrance ramp pole can ensure that the imported high-definition smart camera thereon can take panoramic photos of the front and side of the vehicle entering the area, the license plate and the model; after the photos are taken, they are uploaded to the fusion recognition module for recognition; the number of people in the vehicle entering the area is counted by the imported thermal imaging camera; The exit ramp pole is arranged at a position that ensures that the exit high-definition smart camera thereon can take panoramic photos of the front, license plate and model of the vehicle leaving the area from the front and side; after the photos are taken, they are uploaded to the fusion recognition module for recognition; The exit thermal imaging camera is used to count the number of people in the vehicles leaving the area; the license plate information is the unique identifier of each vehicle and is used to identify the corresponding vehicle; The fusion recognition module includes a vehicle front feature recognition unit, a vehicle side feature recognition unit, a key vehicle track recognition unit and a vehicle interior personnel recognition unit; The vehicle and passenger flow monitoring and judgment module includes a vehicle flow early warning monitoring and judgment unit, a passenger flow early warning monitoring and judgment unit, and a key vehicle electronic fence unit; The warning information release module is used to release warning information; the warning information is released in three ways: ① setting up an electronic display screen and broadcasting equipment 200-500 meters upstream of the service area entrance to release the warning information; ② releasing the warning information through the variable information board on the ETC gantry upstream of the service area; ③ releasing the warning information to the client service module; The client service module is installed on the relevant user's mobile phone and / or car computer, and is used to determine whether to actively pop up the warning information according to the current location of the machine, and passively display the warning information and basic information of the service area according to user instructions.

2. The highway closed service area vehicle and passenger flow accurate monitoring and early warning service system according to claim 1 is characterized in that: The specific process of determining whether to actively pop up the warning information based on the current location of the local machine is that after the client service module receives the warning information issued by the warning information release module, it first determines whether the current navigation route of the local machine passes through the service area, and then determines whether the local machine is within the coverage of the warning information. If both judgment results are yes, the warning information is automatically popped up through the local display device.

3. The precise monitoring and early warning service system for vehicle and passenger flow in closed service areas of expressways according to claim 2 is characterized in that: The determining whether the local machine is within the coverage of the warning information is specifically determining whether the current location of the local machine is within a certain distance upstream of the service area.

4. A method for accurately monitoring and early warning traffic and passenger flow in a closed service area of ​​a highway, characterized in that: The system is implemented based on any one of claims 1 to 3, and specifically comprises the following steps: S1, collecting the traffic information in and out of the closed service area in real time, and calculating the total number of existing vehicles in the service area through a closed service area real-time traffic status monitoring algorithm; The step S1 specifically comprises: calculating the real-time traffic state of the closed service area through the real-time data of the entrance ramp sensing unit and the exit ramp sensing unit; setting the difference in the total number of vehicles entering and leaving the service area measured by the entrance and exit ramp sensing units to the total number of existing vehicles in the service area, and the device numbers of the entrance and exit ramp sensing units of the service area are respectively: the number of the i-th import vehicle, denoted as Cin i , i = 1-N, the number of vehicles at the jth exit, denoted as Cout j , j = 1-M, then the total number of existing vehicles in the service area is expressed by the formula: S2, identifying and acquiring the type of existing vehicles in the service area based on the CNN-based vehicle side feature recognition algorithm according to the data obtained by the entrance and exit ramp sensing unit in step S1; The CNN-based vehicle side feature recognition algorithm of step S2 specifically includes: using the convolutional neural network CNN in deep learning technology as the core model, and reducing the complexity by designing the FlowNet structure; the first is FlowNetHalf, which reduces the number of convolution kernels in each layer of FlowNet to half, and reduces the complexity to 1 / 4. The second is FlowNetInception, which uses the Inception network structure to reduce the complexity to 1 / 8, and automatically learns the method of automatically extracting and identifying the side features of the vehicle in the image through fast image processing, multi-layer convolution and pooling operations, thereby improving the accuracy and efficiency of recognition; S3, using the closed parking lot real-time passenger flow status monitoring algorithm to calculate the maximum number of existing passenger flows; The real-time passenger flow status monitoring algorithm of step S3 specifically includes: calculating the number of people in the existing vehicles in the service area through the data measured by the entrance ramp sensing unit and the exit ramp sensing unit, which is the real-time passenger flow number in the service area; binding according to the license plate information as a unique identifier, when the vehicle is monitored to leave, the number of passengers corresponding to the vehicle is also synchronously reduced, and the maximum number of existing passengers in the service area at time t can be obtained; then the real-time passenger flow status monitoring algorithm is expressed by the formula: Passengers t , represents the current number of passengers in the service area at time t; It represents the jth vehicle of the sensing unit of the import ramp i at time t; represents the number of passengers in the jth vehicle of the entrance / exit ramp sensing unit at time t; It represents the jth vehicle of the ramp sensing unit at exit i at time t; S4, counting the total number of parking spaces of various types in the service area, and then obtaining the total number of all parking spaces; The step S4 specifically includes: the total number of parking spaces in the service area is recorded as Park max , the total number of parking spaces of various types in the service area is counted. The first type is small passenger car parking spaces, denoted as P Car The second category includes parking spaces for large buses and ordinary trucks, denoted by P Bus The third category is parking spaces for large trucks and hazardous chemicals transport vehicles, denoted by P Van ; The fourth category is parking spaces for other types of vehicles, denoted by P other ; then the total number of parking spaces in the service area Park max The calculation formula is: Park max =Σ(P Van +P Car +P Bus +P other ); S5, counting the upper limit of passenger flow capacity of each area in the service area, and then obtaining the maximum upper limit of passenger flow capacity of the service area; The step S5 specifically includes: the maximum passenger flow capacity of the service area is recorded as Passengers max , is the sum of the passenger flow upper limit capacity of the comprehensive business area, public toilets, and other passenger flow areas within the service area. The calculation formula is: Passengers max =Σ(Passengers Business +Passengers toilet +Passengers other ) Where: Passengers Business Indicates the passenger flow capacity limit of the comprehensive business area, Passengers toilet Indicates the upper limit of the passenger flow capacity of the public toilet, Passengers other Indicates the upper limit of passenger flow capacity in other passenger flow areas; S6, calculating the current congestion situation of the service area through a service area early warning analysis algorithm based on the above data; The service area early warning analysis algorithm of step S6 includes: a real-time status algorithm for parking spaces in the service area and a real-time status algorithm for passenger flow in the service area; In the service area parking space real-time status algorithm, the service area parking space real-time status is divided into three states: 0 idle; 1 normal; 2 crowded, and the algorithm is as follows: R Vehicle idle =Car now / Park max , R Vehicle idle <0.6, indicating idle, the real-time status is displayed as 0 idle; 0.6≤R Vehicle idle <1.0, indicating normal, the real-time status is 1 normal, R Vehicle idle ≥1.0, indicating congestion, showing the real-time status as 2 congestion; In the service area passenger flow real-time state algorithm, the service area passenger flow real-time state is divided into three states: 0 idle; 1 normal; 2 crowded, and the algorithm is as follows: R passenger idle =Passengers now / Passengers max , R Passengers idle <0.6, indicating idle, the real-time status is displayed as 0 idle; 0.6≤R Passengers idle <1.0, indicating normal, the real-time status is 1 normal, R Passengersidle ≥1.0, indicating congestion, showing the real-time status as 2 congestion; S7, constructing a vehicle electronic fence for key vehicles, and monitoring the path trajectory and residence time of the key vehicles in the service area, and triggering an early warning and taking preset measures if relevant regulations are violated; The step S7 specifically includes: constructing a vehicle electronic fence in the parking area for hazardous chemicals transport vehicles in the service area, and giving an early warning of the parking position and parking time of hazardous chemicals transport vehicles entering the service area; binding the hazardous chemicals transport vehicle with its license plate information, and tracking the complete path trajectory of the hazardous chemicals transport vehicle in the service area; and using the GPS Beidou positioning data of key operating vehicles for data fitting to realize electronic fence management in the service area. If the hazardous chemicals transport vehicle enters the service area but does not enter the electronic fence corresponding to the parking area for hazardous chemicals transport vehicles, an early warning is issued, or if the vehicle is in the electronic fence for more than half an hour, the service area staff is notified to conduct on-site inquiries, and if it exceeds 1 hour, the owner of the vehicle and the company to which the vehicle belongs are notified to issue an overtime early warning, thereby realizing the full process monitoring of hazardous chemicals transport vehicles from entry to departure, and actively pushing monitoring information to the driver, owner and car company of the vehicle, issuing an overtime early warning and a telephone warning, and ending the full process monitoring when leaving the service area; S8, display the service area warning information board through the electronic display screen and variable information board set up upstream of the service area entrance, and indicate the warning information and basic information of the service area; and prompt relevant users through information push or broadcast.

5. The method for accurate monitoring and early warning of vehicle and passenger flow in a closed service area of ​​a highway according to claim 4 is characterized in that: The step S2 specifically includes: ① Data preprocessing: Preprocess the vehicle side images obtained and transmitted by the entrance and exit ramp perception unit, including image scaling, normalization, and enhancement operations to improve the robustness of the model; ② Feature extraction: Use the pre-trained CNN model ResNet or VGG or a custom CNN model to extract the features of the vehicle side image; these features include the vehicle's outline, windows, and wheels; ③ Feature classification: The extracted features are input into the fully connected layer or classifier to classify the vehicle types; the classifier uses a simple Softmax classifier to classify the vehicle types entering the service area into Among them: 1 small passenger car, 2 large passenger car, 3 ordinary truck, 4 large truck, 5 hazardous chemical transport vehicle, 6 livestock transport vehicle, 7 others; ④ Identify and count the number of people in the car: The thermal imaging camera in the entrance and exit ramp sensing unit takes side photos of the passing vehicles, and uses the deep learning algorithm to identify the number of vehicle occupants and count them. And bind it with the license plate; ⑤ Taking K cars entering the parking lot at time T as an example, the number of passengers on K cars is:

6. The method for accurate monitoring and early warning of vehicle and passenger flow in a closed service area of ​​a highway according to claim 4, characterized in that: The content published and displayed on the electronic display screens and variable information boards should comply with the relevant provisions of the current "Highway Speed ​​Limit Sign Design Specifications" and "Highway LED Variable Information Signs".

Images