High-speed traffic flow prediction and control system and method based on machine vision

Abstract

The present invention relates to the field of high-speed traffic technology, and specifically to a high-speed traffic flow prediction and control system and method based on machine vision; including a camera device, a traffic flow unit and a control strategy unit; the present invention integrates multiple actual influencing factors such as visibility, rainfall, snowfall, wind force and vehicle density, so that the traffic flow finally calculated is more in line with the actual complex traffic environment, and then fully considers the interaction between multiple factors, more accurately reflects the actual traffic carrying capacity and actual traffic flow of each lane under different conditions, and avoids control decision-making errors caused by insufficient consideration of influencing factors; at the same time, it can predict the development trend of traffic flow in advance, improve the foresight and initiative of traffic control, and in addition, it can also realize the refined adjustment of traffic control strategies and improve the operating efficiency of the entire transportation system.

Description

Technical Field

[0001] The present invention relates to the field of high-speed traffic technology, and in particular to a high-speed traffic flow prediction and control system and method based on machine vision. Background Art

[0002] With the rapid development of the economy, the traffic flow on highways has shown a trend of sustained growth, and the problem of traffic congestion has become increasingly serious, bringing many inconveniences to people's travel and cargo transportation, while also causing a series of social problems such as energy waste and environmental pollution.

[0003] In modern traffic management concepts, predicting traffic flow in advance and implementing intelligent control based on the prediction results have become the key to improving traffic operation efficiency and alleviating congestion. Traffic flow prediction aims to analyze historical traffic data and combine real-time monitoring information to explore the changing patterns of traffic flow, and to predict the changes in parameters such as traffic flow, speed, density, etc. in a certain period of time in the future in advance, so that traffic management departments can prepare for response in advance. However, the current traffic conditions are affected by many factors, such as weather, holidays, emergencies, etc., making short-term traffic flow prediction difficult.

[0004] The present invention provides a high-speed traffic flow prediction and control system and method based on machine vision to solve the above-mentioned technical problems. Summary of the invention

[0005] The purpose of the present invention is to provide a high-speed traffic flow prediction and control system and method based on machine vision, which integrates multiple actual influencing factors such as visibility, rainfall, snowfall, wind force and vehicle density, so that the traffic flow finally calculated is more in line with the actual complex traffic environment, and then fully considers the interaction between multiple factors, and more accurately reflects the actual traffic carrying capacity and actual traffic flow of each lane under different conditions, avoiding control decision-making errors caused by insufficient consideration of influencing factors; at the same time, it can predict the development trend of traffic flow in advance, improve the foresight and initiative of traffic control, and in addition, it can also realize the refined adjustment of traffic control strategies and improve the operating efficiency of the entire transportation system.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] The first aspect of the present invention provides a high-speed traffic flow prediction and control system based on machine vision, comprising a camera device, a traffic flow unit and a control strategy unit, wherein:

[0008] The camera device is used to collect traffic images of the target road in real time;

[0009] The traffic flow unit is used to extract traffic image features from the obtained traffic image, calculate and obtain traffic flow data, and then obtain the traffic flow of the corresponding lane according to the traffic flow data and the traffic image features, and obtain the predicted traffic flow data using the constructed traffic flow prediction model, and the traffic flow unit is connected to the camera device;

[0010] The control strategy unit controls the target road according to the acquired predicted traffic flow data, uploads the control strategy and retrieves the information, and the control strategy unit is connected to the traffic flow unit.

[0011] The present invention is further configured as follows: the camera device is a camera with a wireless communication function.

[0012] The present invention is further configured as follows: the traffic flow unit includes a first communication module, an image feature module and a traffic flow module, wherein:

[0013] The first communication module is used to realize information interaction between the traffic flow unit and the camera device and the control strategy unit;

[0014] The image feature module extracts corresponding traffic image features from the acquired traffic images respectively with the monitoring period as the extraction interval, and the image feature module is connected to the first communication module;

[0015] The traffic flow module calculates and obtains traffic flow data according to the acquired traffic image features, and then obtains the traffic flow of the corresponding lane according to the traffic flow data and the traffic image features. The traffic flow module is connected to the image feature module.

[0016] The present invention is further configured as follows: the traffic flow unit further includes a flow prediction module and a database module, wherein:

[0017] The traffic prediction module is used to construct a traffic volume prediction model and obtain predicted traffic flow data by taking the traffic flow data acquired in the current continuous monitoring cycle as input, and the traffic prediction module is connected to the traffic flow module;

[0018] The database module is used to store the trained traffic volume prediction model, as well as the received traffic images and predicted traffic flow data.

[0019] The present invention is further configured as follows: the control strategy unit includes a second communication module, a traffic comparison module, a strategy storage module, an information upload module and an information retrieval module, wherein:

[0020] The second communication module is used to realize information interaction between the control strategy unit and the traffic flow unit;

[0021] The traffic flow comparison module compares the predicted traffic flow data with the preset traffic flow control data to control the target road, and the traffic flow comparison module is connected to the second communication module;

[0022] The policy storage module is used to store preset control policies, and the policy storage module is connected to the traffic comparison module;

[0023] The information uploading module is used to upload the preset management and control strategy, and the information uploading module is connected to the strategy storage module;

[0024] The information retrieval module is used to retrieve the stored traffic images and predicted traffic flow data, and the information retrieval module is connected to the second communication module.

[0025] The second aspect of the present invention is to provide a method for predicting and controlling high-speed traffic flow based on machine vision, which uses the above-mentioned high-speed traffic flow prediction and control system based on machine vision, and comprises the following steps:

[0026] Use camera equipment to collect traffic images of target roads in real time;

[0027] Taking the monitoring period as the extraction interval, the corresponding traffic image features are extracted from the obtained traffic images respectively;

[0028] Calculate and obtain traffic flow data according to the acquired traffic image features;

[0029] According to the traffic flow data and traffic image features, the traffic flow of the corresponding lane is obtained;

[0030] Obtain the historical continuous traffic flow data of the corresponding lanes of the target road, train the convolutional neural network, and obtain the traffic flow prediction model;

[0031] Using the traffic flow data acquired in the current continuous monitoring period as input, inputting into the traffic flow prediction model, and obtaining the predicted traffic flow data;

[0032] The target road is controlled based on the predicted traffic flow data obtained.

[0033] The present invention is further configured as follows: the process of extracting the traffic image features is as follows:

[0034] Taking the starting point and the end point of the monitoring period as the time points for acquiring the traffic images, acquiring a first traffic image at the starting point and a second traffic image at the end point;

[0035] Extracting first traffic image features and second traffic image features from the first traffic image and the second traffic image respectively;

[0036] The first traffic image feature and the second traffic image feature both include the pixel position of the vehicle in the traffic image, the number of lanes, the number of vehicles in the corresponding lanes, and the environmental information in the corresponding monitoring period;

[0037] The environmental information includes visibility, rainfall, snowfall, and wind information.

[0038] The present invention is further configured as follows: the traffic flow data includes vehicle speed and vehicle density, wherein:

[0039] Vehicle speed: The speed of each vehicle is calculated based on the displacement difference between the first traffic image feature and the second traffic image feature during the monitoring period. Where x is the number of pixels moved horizontally by vehicle i during the monitoring period; y is the number of pixels moved vertically by vehicle i during the monitoring period, and Δt is the monitoring period. Then the average speed of vehicles in each lane is calculated. Where N is the number of vehicles in the corresponding lane;

[0040] Vehicle density: Calculate the vehicle density of the corresponding lane based on the number of vehicles in the monitoring area Where N j is the number of vehicles in the corresponding lane j, L j is the length of the corresponding lane j in the monitoring area.

[0041] The present invention is further configured as follows: the calculation formula for the traffic flow of the corresponding lane is: in:

[0042] q0 is the traffic capacity of the lane under ideal conditions;

[0043] Where V is the actual visibility, V o is the critical visibility threshold, n is the empirical coefficient;

[0044] f(R)=(1-k R ×R m ), where k R is the rainfall influence coefficient, R is the rainfall intensity, and m is the empirical index;

[0045] Where S is the snowfall, k S is the snowfall influence coefficient;

[0046] Where D is the wind force level, D max is the maximum wind force level that has a significant impact on traffic, and p is an empirical parameter;

[0047] In the formula, ρj is the blocking density.

[0048] The present invention is further configured as follows: the process of controlling the target road is as follows:

[0049] If the predicted traffic flow data is greater than the preset traffic flow control data, the preset control strategy for the corresponding target road is activated;

[0050] Otherwise, the normalized control strategy will be used as the control strategy for the target road.

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] The present invention collects images through machine vision technology and calculates vehicle speed, vehicle density and traffic flow based on them, which can cover all lanes and many vehicles in the monitoring area. By integrating multiple actual influencing factors such as visibility, rainfall, snowfall, wind and vehicle density, the traffic flow finally calculated is more in line with the actual complex traffic environment, and then fully considers the interaction between multiple factors, more accurately reflects the real traffic carrying capacity and actual traffic flow of each lane under different conditions, and avoids the control decision-making errors caused by insufficient consideration of influencing factors; at the same time, because it is based on continuous traffic images for analysis, not only can the traffic flow data at a certain moment be obtained, but also the change process of each parameter of the traffic flow can be continuously tracked over time, so as to predict the development trend of the traffic flow in advance, and improve the foresight and initiative of traffic control. In addition, based on the data such as traffic flow that comprehensively considers multiple factors, combined with the comparison of the prediction results with the preset control data, the fine adjustment of the traffic control strategy can be realized, which meets the actual traffic conditions of each lane, different weather conditions and real-time dynamic requirements of traffic flow changes, accurately selects appropriate control means, maximizes the road capacity, reduces the probability of traffic congestion, and improves the operation efficiency of the entire transportation system. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 This is a system diagram of the high-speed traffic flow prediction and control system based on machine vision of the present invention.

[0054] Figure 2 This is a system diagram of the traffic flow unit in the high-speed traffic flow prediction and control system based on machine vision of the present invention.

[0055] Figure 3 This is a system diagram of the control strategy unit in the high-speed traffic flow prediction and control system based on machine vision of the present invention.

[0056] Legend: 100, camera device; 200, traffic flow unit; 210, first communication module; 220, image feature module; 230, traffic flow module; 240, traffic prediction module; 250, database module; 300, management and control strategy unit; 310, second communication module; 320, traffic comparison module; 330, strategy storage module; 340, information upload module; 350, information retrieval module. DETAILED DESCRIPTION

[0057] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0058] Example:

[0059] like Figure 1-Figure 3 As shown, this embodiment provides a high-speed traffic flow prediction and control system based on machine vision, including a camera device 100, a traffic flow unit 200 and a control strategy unit 300, wherein: the camera device 100 is used to collect traffic images of the target road in real time; the traffic flow unit 200 is used to extract traffic image features from the obtained traffic images, and calculate and obtain traffic flow data, and then obtain the traffic flow of the corresponding lane according to the traffic flow data and the traffic image features, and obtain the predicted traffic flow data by using the constructed traffic flow prediction model, and the traffic flow unit 200 is connected to the camera device 100; the control strategy unit 300 controls the target road according to the obtained predicted traffic flow data, and uploads the control strategy and retrieves the information, and the control strategy unit 300 is connected to the traffic flow unit 200.

[0060] The camera device 100 is a camera with a wireless communication function.

[0061] In the present embodiment, it should be noted that the traffic image of the target road is obtained by the provided camera device 100, and the traffic image of the target road is uploaded to the traffic flow unit 200. The traffic flow unit 200 first extracts the traffic image features from the obtained traffic image, and calculates and obtains the traffic flow data. Then, the traffic flow of the corresponding lane is obtained according to the traffic flow data and the traffic image features, and the predicted traffic flow data is obtained by using the constructed traffic flow prediction model. The pixel position of the vehicle in the traffic image, the number of lanes, the number of vehicles in the corresponding lane and the environmental information in the corresponding monitoring period can be integrated to obtain the traffic flow integrating the environmental information. Compared with the simple traffic flow calculation based on the number of vehicles passing through, it is more scientific and accurate, and can reflect the real situation of traffic flow in the actual complex environment, laying the foundation for the subsequent accurate prediction and effective control. At the same time, the predicted traffic flow data will be transmitted to the control strategy unit 300, and the control strategy unit 300 will perform control, that is, the control strategy adopted for the target road is determined according to the comparison result between the predicted traffic flow data and the preset traffic flow control data. If the predicted data is greater than the preset control data, the preset control strategy of the corresponding target road will be activated (such as implementing temporary speed limits, adjusting lane functions, etc.); otherwise, the normalized control strategy will be adopted. This control method based on comparative judgment can realize dynamic and intelligent control of traffic flow, respond to changes in traffic flow in a timely manner, and ensure smooth and safe traffic on the target road.

[0062] In the present invention, the traffic flow unit 200 includes a first communication module 210, an image feature module 220 and a traffic flow module 230, wherein: the first communication module 210 is used to realize information interaction between the traffic flow unit 200 and the camera device 100 and the control strategy unit 300; the image feature module 220 extracts corresponding traffic image features from the obtained traffic image with the monitoring period as the extraction interval, and the image feature module 220 is connected to the first communication module 210; the traffic flow module 230 calculates and obtains traffic flow data based on the obtained traffic image features, and then obtains the traffic flow of the corresponding lane based on the traffic flow data and the traffic image features, and the traffic flow module 230 is connected to the image feature module 220.

[0063] In addition, the traffic flow unit 200 further includes a traffic flow prediction module 240 and a database module 250, wherein:

[0064] The traffic prediction module 240 is used to construct a traffic volume prediction model, and to obtain predicted traffic flow data using the traffic flow data acquired in the current continuous monitoring period as input. The traffic prediction module 240 is connected to the traffic flow module 230; the database module 250 is used to store the trained traffic volume prediction model, as well as the received traffic images and predicted traffic flow data.

[0065] In the present embodiment, it should be noted that the information of the camera device 100 is received by the first communication module 210, and the above information is transmitted to the image feature module 220. The image feature module 220 obtains the corresponding traffic image features, and transmits the extracted traffic image features to the traffic flow module 230, calculates and obtains the traffic flow data, and then obtains the traffic flow of the corresponding lane based on the traffic flow data and the traffic image features, and the traffic flow data is predicted by the traffic prediction module 240. In the present embodiment, the monitoring period is used as the extraction interval to obtain the traffic images of the starting point and the end point, and extracts multi-dimensional traffic image features including the vehicle pixel position, the number of lanes, the number of vehicles in the lane, and environmental information.

[0066] In the present invention, the control strategy unit 300 includes a second communication module 310, a traffic comparison module 320, a strategy storage module 330, an information upload module 340 and an information retrieval module 350, wherein:

[0067] The second communication module 310 is used to realize information interaction between the control strategy unit 300 and the traffic flow unit 200; the traffic comparison module 320 compares the predicted traffic flow data with the preset traffic flow control data to control the target road, and the traffic comparison module 320 is connected to the second communication module 310; the strategy storage module 330 is used to store the preset control strategy, and the strategy storage module 330 is connected to the traffic comparison module 320; the information upload module 340 is used to upload the preset control strategy, and the information upload module 340 is connected to the strategy storage module 330; the information retrieval module 350 is used to retrieve the stored traffic images and predicted traffic flow data, and the information retrieval module 350 is connected to the second communication module 310.

[0068] In this embodiment, it should be noted that the second communication module 310 receives the predicted traffic flow data of the traffic flow unit 200, and then the traffic comparison module 320 compares it with the preset traffic flow control data to control the target road. The management personnel can upload the preset control strategy through the information upload module 340 and store it in the strategy storage module 330. When the traffic comparison module 320 performs the comparison, the corresponding preset traffic flow control data and control strategy can be retrieved from the strategy storage module 330.

[0069] In addition, this embodiment also provides a high-speed traffic flow prediction and control method based on machine vision, which uses the above-mentioned high-speed traffic flow prediction and control system based on machine vision, including the following steps:

[0070] The camera equipment is used to collect traffic images of the target road in real time.

[0071] The monitoring period is used as the extraction interval, and the corresponding traffic image features are extracted from the obtained traffic images.

[0072] Among them, the process of extracting traffic image features is as follows:

[0073] Taking the starting point and the end point of the monitoring period as the traffic image acquisition time points, acquiring a first traffic image at the starting point and a second traffic image at the end point; extracting first traffic image features and second traffic image features from the first traffic image and the second traffic image respectively;

[0074] Furthermore, the first traffic image feature and the second traffic image feature both include the pixel position of the vehicle in the traffic image, the number of lanes, the number of vehicles in the corresponding lanes, and environmental information in the corresponding monitoring period; the environmental information includes visibility, rainfall, snowfall, and wind information.

[0075] Traffic flow data is calculated and acquired based on the acquired traffic image features.

[0076] The traffic flow data includes vehicle speed and vehicle density, where:

[0077] Vehicle speed: The speed of each vehicle is calculated based on the displacement difference between the first traffic image feature and the second traffic image feature during the monitoring period. Where x is the number of pixels moved horizontally by vehicle i during the monitoring period; y is the number of pixels moved vertically by vehicle i during the monitoring period, and Δt is the monitoring period. Then the average speed of vehicles in each lane is calculated. Where N is the number of vehicles in the corresponding lane;

[0078] Vehicle density: Calculate the vehicle density of the corresponding lane based on the number of vehicles in the monitoring area Where N j is the number of vehicles in the corresponding lane j, L j is the length of the corresponding lane j in the monitoring area.

[0079] The present invention is further configured as follows: the calculation formula for the traffic flow of the corresponding lane is: in:

[0080] q0 is the traffic capacity of the lane under ideal conditions;

[0081] Where V is the actual visibility, V o is the critical visibility threshold, n is the empirical coefficient;

[0082] f(R)=(1-k R ×R m ), where k R is the rainfall influence coefficient, R is the rainfall intensity, and m is the empirical index;

[0083] Where S is the snowfall, k S is the snowfall influence coefficient;

[0084] Where D is the wind force level, D max is the maximum wind force level that has a significant impact on traffic, and p is an empirical parameter;

[0085] In the formula, ρ j is the blocking density.

[0086] According to the traffic flow data and traffic image features, the traffic flow of the corresponding lane is obtained.

[0087] The historical continuous traffic flow data of the corresponding lanes of the target road is obtained, and the convolutional neural network is trained to obtain the traffic flow prediction model.

[0088] The traffic flow data obtained in the current continuous monitoring period is used as input and input into the traffic flow prediction model to obtain the predicted traffic flow data.

[0089] The target road is controlled based on the predicted traffic flow data obtained.

[0090] The process of controlling the target road is as follows:

[0091] If the predicted traffic flow data is greater than the preset traffic flow control data, the preset control strategy for the corresponding target road will be activated; otherwise, the normalized control strategy will be used as the control strategy for the target road.

[0092] In this embodiment, it should be noted that multi-dimensional traffic image features are extracted from the traffic image at intervals of the monitoring period, including the pixel position of the vehicle in the image, the number of lanes, the number of vehicles in the corresponding lanes, and environmental information (visibility, rainfall, snowfall, wind information, etc.); the vehicle pixel position information can be used to subsequently accurately calculate the displacement of the vehicle and then derive the vehicle speed; the number of lanes clarifies the basic configuration of traffic resources, which helps to perform traffic flow analysis for different lanes; the number of vehicles in the lane is directly related to the calculation of traffic flow; and the environmental information takes into account the impact of external factors on traffic flow, so that the subsequent calculation and analysis of traffic flow data are more in line with the actual situation.

[0093] The provided method comprehensively considers the lane capacity under ideal conditions as well as multiple actual influencing factors such as visibility, rainfall, snowfall, wind, vehicle density, etc., and uses a specific formula to obtain the traffic flow of the corresponding lane. It can fully integrate the relationship between environmental factors and traffic flow, and more truly and accurately reflect the traffic carrying capacity and actual capacity of each lane under actual complex road conditions.

[0094] By comparing the predicted traffic flow data with the preset traffic flow control data, the control strategy for the target road is determined based on the comparison results. If the predicted data is greater than the preset control data, it means that the traffic flow is about to exceed the normal control range. At this time, the preset control strategy of the corresponding target road is activated (such as adjusting the lane function, implementing variable speed limits, controlling the entry of ramp vehicles, etc.), and timely intervention and guidance of traffic flow is carried out to avoid the formation or deterioration of congestion; if the predicted data does not exceed the preset value, the normalized control strategy is adopted to maintain the normal operation of traffic, so that the control measures can be flexibly adjusted according to the real-time change trend of traffic flow, and the refined and intelligent management of traffic flow can be realized, so as to maximize the road traffic efficiency and traffic safety.

[0095] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0096] The preferred embodiments of the present invention disclosed above are only used to help illustrate the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A high-speed traffic flow prediction and control system based on machine vision, characterized by: The system comprises a camera device (100), a traffic flow unit (200) and a control strategy unit (300), wherein: The camera device (100) is used to collect traffic images of a target road in real time; The traffic flow unit (200) is used to extract traffic image features from the acquired traffic image, calculate and obtain traffic flow data, and then obtain the traffic flow of the corresponding lane according to the traffic flow data and the traffic image features, and obtain the predicted traffic flow data using the constructed traffic flow prediction model. The traffic flow unit (200) is connected to the camera device (100); The control strategy unit (300) controls the target road according to the acquired predicted traffic flow data, uploads the control strategy and retrieves information. The control strategy unit (300) is connected to the traffic flow unit (200).

2. The high-speed traffic flow prediction and control system based on machine vision according to claim 1 is characterized in that: The camera device (100) is a camera with a wireless communication function.

3. The high-speed traffic flow prediction and control system based on machine vision according to claim 1 is characterized in that: The traffic flow unit (200) comprises a first communication module (210), an image feature module (220) and a traffic flow module (230), wherein: The first communication module (210) is used to realize information interaction between the traffic flow unit (200) and the camera device (110) and the control strategy unit (300); The image feature module (220) extracts corresponding traffic image features from the acquired traffic images respectively with the monitoring period as the extraction interval, and the image feature module (220) is connected to the first communication module (210); The traffic flow module (230) calculates and obtains traffic flow data based on the acquired traffic image features, and then obtains the traffic flow of the corresponding lane based on the traffic flow data and the traffic image features. The traffic flow module (230) is connected to the image feature module (220).

4. The high-speed traffic flow prediction and control system based on machine vision according to claim 3 is characterized in that: The traffic flow unit (200) further comprises a flow prediction module (240) and a database module (250), wherein: The traffic prediction module (240) is used to construct a traffic volume prediction model and to obtain predicted traffic flow data using traffic flow data acquired in the current continuous monitoring cycle as input, and the traffic prediction module (240) is connected to the traffic flow module (230); The database module (250) is used to store the trained traffic volume prediction model, as well as the received traffic images and predicted traffic flow data.

5. The high-speed traffic flow prediction and control system based on machine vision according to claim 1 is characterized in that: The control strategy unit (300) comprises a second communication module (310), a traffic comparison module (320), a strategy storage module (330), an information upload module (340) and an information retrieval module (350), wherein: The second communication module (310) is used to realize information interaction between the control strategy unit (300) and the traffic flow unit (200); The traffic flow comparison module (320) compares the predicted traffic flow data with the preset traffic flow control data to control the target road, and the traffic flow comparison module (320) is connected to the second communication module (310); The strategy storage module (330) is used to store preset management and control strategies, and the strategy storage module (330) is connected to the flow comparison module (320); The information uploading module (340) is used to upload the preset management and control strategy, and the information uploading module (340) is connected to the strategy storage module (330); The information retrieval module (350) is used to retrieve the stored traffic images and predicted traffic flow data, and the information retrieval module (350) is connected to the second communication module (310).

6. A high-speed traffic flow prediction and control method based on machine vision, applied to a high-speed traffic flow prediction and control system based on machine vision as claimed in any one of claims 1 to 5, characterized in that: The following steps are involved: Use camera equipment to collect traffic images of target roads in real time; Taking the monitoring period as the extraction interval, the corresponding traffic image features are extracted from the obtained traffic images respectively; Calculate and obtain traffic flow data according to the acquired traffic image features; According to the traffic flow data and traffic image features, the traffic flow of the corresponding lane is obtained; Obtain the historical continuous traffic flow data of the corresponding lanes of the target road, train the convolutional neural network, and obtain the traffic flow prediction model; Using the traffic flow data acquired in the current continuous monitoring period as input, inputting into the traffic flow prediction model, and obtaining the predicted traffic flow data; The target road is controlled based on the predicted traffic flow data obtained.

7. The high-speed traffic flow prediction and control method based on machine vision according to claim 5 is characterized in that: The process of extracting the traffic image features is as follows: Taking the starting point and the end point of the monitoring period as the time points for acquiring the traffic images, acquiring a first traffic image at the starting point and a second traffic image at the end point; Extracting first traffic image features and second traffic image features from the first traffic image and the second traffic image respectively; The first traffic image feature and the second traffic image feature both include the pixel position of the vehicle in the traffic image, the number of lanes, the number of vehicles in the corresponding lanes, and the environmental information in the corresponding monitoring period; The environmental information includes visibility, rainfall, snowfall, and wind information.

8. The high-speed traffic flow prediction and control method based on machine vision according to claim 7 is characterized in that: The traffic flow data includes vehicle speed and vehicle density, wherein: Vehicle speed: The speed of each vehicle is calculated based on the displacement difference between the first traffic image feature and the second traffic image feature during the monitoring period. Where x is the number of pixels moved horizontally by vehicle i during the monitoring period; y is the number of pixels moved vertically by vehicle i during the monitoring period, and Δt is the monitoring period. Then the average speed of vehicles in each lane is calculated. Where N is the number of vehicles in the corresponding lane; Vehicle density: Calculate the vehicle density of the corresponding lane based on the number of vehicles in the monitoring area Where N j is the number of vehicles in the corresponding lane j, L j is the length of the corresponding lane j in the monitoring area.

9. The high-speed traffic flow prediction and control method based on machine vision according to claim 5 is characterized in that: The calculation formula for the traffic flow of the corresponding lane is: in: q0 is the traffic capacity of the lane under ideal conditions; Where V is the actual visibility, V o is the critical visibility threshold, n is the empirical coefficient; f(R)=(1-k R ×R m ), where k R is the rainfall influence coefficient, R is the rainfall intensity, and m is the empirical index; Where S is the snowfall, k S is the snowfall influence coefficient; Where D is the wind force level, D max is the maximum wind force level that has a significant impact on traffic, and p is an empirical parameter; In the formula, ρ j is the blocking density.

10. The high-speed traffic flow prediction and control method based on machine vision according to claim 5, characterized in that: The process of controlling the target road is as follows: If the predicted traffic flow data is greater than the preset traffic flow control data, the preset control strategy for the corresponding target road is activated; Otherwise, the normalized control strategy will be used as the control strategy for the target road.

Images