Highway safety early warning method and device based on machine vision, storage medium and computer program product

Abstract

The invention discloses an expressway safety early warning method and device based on machine vision, a storage medium and a computer program product, and belongs to the field of construction safety. According to the method, the accuracy of highway safety early warning can be effectively improved by performing segmented evaluation on the to-be-evaluated segments, and the segmentation of the segments can be performed according to the recognition efficiency of a machine vision algorithm, so that the recognition efficiency of the algorithm on the information of each segment after the segmentation is ensured. And after the risk evaluation thresholds are delimited, determining the risk evaluation level of the evaluation section of the expressway by determining the position of the second risk evaluation value among the plurality of risk evaluation thresholds. And finally, the safety risk levels of different risk monitoring sections of the expressway can be displayed through a safety risk management map, and different risk levels correspond to different colors, so that management and control personnel can check the safety risk of the expressway more visually.

Description

Technical Field

[0001] The present application relates to the field of construction safety, and in particular to a highway safety warning method, device, storage medium and computer program product based on machine vision. Background Art

[0002] At a time when artificial intelligence and autonomous driving technologies are advancing rapidly, automation and intelligence have become necessary means for highway construction safety management. Solving the safety protection issues of drivers on highway sites through information technology has become a trend of the times. For this reason, the highway safety warning system was born. With the help of early warning devices, traffic safety hazards can be checked, and safety risk items in the operation of highways can be warned and prevented in a timely manner. In addition, through the early warning function of the system, it can assist drivers and operation dispatchers in solving key factors in the driving process, such as vehicles, pedestrians, road conditions, natural environment and other safety hazard information, and promptly remind drivers and dispatchers to take corresponding safety prevention and control measures, thereby improving the danger response speed and crisis response capabilities of highway users.

[0003] Especially when it comes to road construction safety, there are safety risks such as lane fences blocking the line of sight, limited road width and route width, and strong random distribution of personnel. If vehicles change lanes, the safety risks during highway operation will increase exponentially. At present, there are some cases of solving construction safety risk monitoring through machine vision algorithms. However, the relevant cases are mainly static monitoring, which cannot effectively solve the safety risks caused by high-speed vehicles on highways, sudden changes in natural risks, and irregular distribution of construction personnel.

[0004] In addition, some methods overly focus on accuracy and real-time performance, while ignoring the complexity and lightweight requirements of the equipment, which is not conducive to construction sites that require high portability. The current systems are mostly suitable for multi-lane coordinated warnings, but in the construction of existing lines, one-way or two-way roads are more common, which reduces the accuracy of the system's warnings. Most systems rely on only a single prediction model as a judgment criterion, and the accuracy is insufficient for complex road sections. Summary of the invention

[0005] The embodiments of the present application provide a highway safety warning method, device, storage medium and computer program product based on machine vision. In order to solve the problem that the application scenarios of highway safety warning methods in the prior art are mainly static monitoring, and cannot effectively solve the safety risks caused by high-speed driving of vehicles on highways, sudden changes in natural risks, irregular distribution of construction personnel, etc., the technical solution is as follows:

[0006] On the one hand, a highway safety warning method based on machine vision is provided, which obtains a monitoring section of the highway along the highway to be detected, and performs safety risk identification and warning on the detection section, including:

[0007] Step S100, dividing the expressway along the route into sections, and dividing the expressway along the route into a plurality of sections to be evaluated;

[0008] Step S200, obtaining risk assessment information of the plurality of sections to be assessed, and determining first risk assessment values ​​corresponding to the plurality of sections to be assessed according to the risk assessment information;

[0009] Step S300: determining second risk assessment values ​​corresponding to the plurality of sections to be assessed based on the first risk assessment value and the risk assessment information;

[0010] Step S400: defining a risk assessment threshold value, which is used to determine a security risk assessment level corresponding to the second risk assessment value;

[0011] Step S500: generating a security risk management map according to the security risk assessment level;

[0012] The risk assessment information includes: vehicle information, personnel location information, natural environment information and temporary operation equipment information in the section to be assessed;

[0013] Optionally, the obtaining risk assessment information of the plurality of sections to be assessed includes:

[0014] Step S201, obtaining vehicle information, the vehicle information including: vehicle speed information, vehicle quantity information, and vehicle congestion information;

[0015] By means of monitoring equipment arranged in the monitoring section of the expressway, photos of vehicles traveling in the monitoring section of the expressway are obtained, and the vehicle quantity information and the vehicle congestion information are identified based on a machine vision algorithm, and the information is quantified and outputted;

[0016] The vehicle position information in the monitoring section of the expressway is monitored in real time by a global positioning system (GPS) deployed in the monitoring section of the expressway, and the vehicle speed information is determined by parameters related to the change of the vehicle position information and the change of time;

[0017] Step S202: obtaining personnel location information, identifying personnel features based on machine vision algorithms through monitoring equipment deployed in the monitoring section of the expressway, and obtaining personnel information in the monitoring section of the expressway;

[0018] Step S203, obtaining natural environment information, and obtaining temperature information, humidity information, wind speed information and precipitation information of the section to be evaluated through the cooperation of a weather station, a temperature sensor, a humidity sensor and a wind speed sensor;

[0019] Step S204: obtain temporary operation equipment information, and monitor the operation information of the temporary operation equipment through the Internet of Things technology.

[0020] Optionally, the identifying the vehicle quantity information and the vehicle congestion information based on a machine vision algorithm and quantifying and outputting the information includes:

[0021] Image acquisition, using surveillance cameras and drone equipment to collect real-time image data within the monitoring section of the expressway;

[0022] Image preprocessing, adjusting the brightness and contrast of the image data, and performing denoising on the image data;

[0023] Vehicle detection, applying a target detection algorithm to identify a vehicle in an image, locate the position of the vehicle in the image, and record the position information of the vehicle;

[0024] Vehicle tracking, using the optical flow method to track vehicles in consecutive frames and identify vehicles with the same features;

[0025] Vehicle quantity counting, counting the vehicles in the image to obtain vehicle quantity information;

[0026] Vehicle congestion analysis, evaluating vehicle congestion based on the distance between vehicles and lane occupancy in the image;

[0027] Congestion level definition, setting congestion evaluation thresholds, and determining different congestion levels corresponding to different vehicle-to-vehicle distances and lane occupancy rates. The congestion levels include: loose, normal, congested, and very congested;

[0028] Quantitative output, converting the vehicle quantity information and the congestion information into specific quantitative index output.

[0029] Optionally, determining, according to the risk assessment information, first risk assessment values ​​corresponding to the plurality of sections to be assessed comprises: constructing a first risk assessment indicator system corresponding to the first risk assessment value by a cluster analysis method, the first risk assessment system comprising vehicle flow, accident rate and environmental factors;

[0030] Determine the corresponding relationship between the risk assessment information and the vehicle flow, the accident rate, and the environmental factors.

[0031] Optionally, determining the second risk assessment value corresponding to the multiple sections to be evaluated includes: based on the environmental factors, determining a weight change table of the first risk assessment index system under different environments, weightedly summing the vehicle flow, the accident rate and the environmental factors to determine the second risk assessment value.

[0032] Optionally, risk assessment thresholds may include:

[0033] Step S401, collecting and analyzing relevant historical data, including accident reports, traffic flow, and environmental conditions;

[0034] Step S402: Divide data features through historical data analysis to determine the initial setting of thresholds;

[0035] Step S403: verifying and adjusting the threshold through expert review and field testing;

[0036] Step S404: Set a threshold adjustment mechanism and record the threshold setting process and reasons.

[0037] Optionally, generating a security risk management map according to the security risk assessment level includes:

[0038] Step S501, combining the second risk assessment value with the GIS platform, and utilizing the spatial data processing and analysis capabilities of GIS;

[0039] Step S502: assign a risk level to each section to be evaluated in the GIS, and mark it with different colors or symbols according to the safety risk evaluation level;

[0040] Step S503: Using the layer function of GIS, additional information layers such as traffic flow, weather conditions, topography, etc. are superimposed to provide a more comprehensive visual analysis;

[0041] Step S504: using a data visualization tool to create an interactive map;

[0042] Step S505: designing visualization elements to represent risk levels;

[0043] Step S506: Implementing zooming, scrolling and click interaction functions of the map so that the user can explore more detailed information;

[0044] Step S507: Integrate the real-time data stream, update the vehicle flow information, accident rate information and environmental factor information, and dynamically generate the safety risk management map;

[0045] Step S508, setting data refresh frequency;

[0046] Step S509: defining a color code as a quantitative indicator for each risk level in the security risk assessment level;

[0047] Step S510: Display the risk level of each section on the map in the form of quantitative indicators;

[0048] Step S511, integrating an early warning system on the map, and issuing a warning for a section with a risk greater than a limit threshold, wherein the limit threshold is the highest value among the risk assessment values ​​of the road being in a safe state;

[0049] Step S512: Design a sound warning signal and notify the on-site management personnel via SMS and phone call.

[0050] On the other hand, a highway safety warning device based on machine vision is provided, comprising a processor, a memory and a computer program / instructions stored in the memory, wherein the processor is used to execute the computer program / instructions, and when the computer program / instructions are executed, the management system implements the steps of the above method.

[0051] On the other hand, a computer-readable storage medium is provided, on which a computer program / instruction is stored, and when the computer program / instruction is executed by a processor, the steps of the above method are implemented.

[0052] On the other hand, a computer program product is provided, comprising a computer program / instruction, characterized in that the computer program / instruction implements the steps of the above method when executed by a processor.

[0053] The beneficial effects brought by the technical solution provided by the embodiment of the present application include at least:

[0054] The embodiment of the present application provides a highway safety warning method based on machine vision. By performing segment evaluation on multiple sections to be evaluated, the accuracy of highway safety warning can be effectively improved. The segment division can be performed based on the recognition efficiency of the machine vision algorithm, thereby ensuring the recognition efficiency of the algorithm for each section information after the segment division. Through the first risk evaluation value, various risk information in the section to be evaluated can be classified and initially evaluated, which is convenient for relevant managers to control the vehicle safety situation of the highway. The second risk evaluation value generated based on the first risk evaluation value is used to perform a fixed value evaluation on the safety risk of the highway. The second risk evaluation value can more accurately evaluate the safety risk situation of the highway and provide sufficient basis for the subsequent safety risk display. After the risk evaluation threshold is delineated, the risk evaluation level of the evaluation section of the highway is determined by determining the position of the second risk evaluation value between multiple risk evaluation thresholds. Finally, the safety risk level of different risk monitoring sections of the highway can be displayed through the safety risk management map, and different risk levels correspond to different colors, so that the management and control personnel can view the safety risk of the highway more intuitively. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0056] Figure 1 It is a flowchart of a highway safety warning method based on machine vision provided in an embodiment of the present application;

[0057] Figure 2 is a schematic diagram of a security risk management map provided in an embodiment of the present application;

[0058] Figure 3 It is a flowchart of a method for obtaining risk assessment information of multiple sections to be assessed provided in an embodiment of the present application. DETAILED DESCRIPTION

[0059] In order to make the objectives, technical solutions and advantages of the present application clearer, the implementation methods of the present application will be further described in detail below with reference to the accompanying drawings.

[0060] Please refer to Figure 1 , Figure 1 The present invention provides a flowchart of a highway safety warning method based on machine vision, which is provided in an embodiment of the present invention. The highway safety warning method based on machine vision obtains a monitoring section of a highway along the line to be detected, identifies safety risks and warns the detected section, and includes:

[0061] Step S100: divide the expressway along the route into sections, and divide the expressway along the route into multiple sections to be evaluated. By performing section-by-section evaluation on multiple sections to be evaluated, the accuracy of expressway safety warning can be effectively improved. The section division can be performed based on the recognition efficiency of the machine vision algorithm, thereby ensuring the recognition efficiency of the algorithm for each section information after the section division.

[0062] Step S200, obtaining risk assessment information of multiple sections to be evaluated, and determining first risk assessment values ​​corresponding to multiple sections to be evaluated according to the risk assessment information. Through the first risk assessment value, various risk information in the section to be evaluated can be classified and initially evaluated, so as to facilitate relevant management personnel to control the vehicle safety situation on the highway. Among them, the risk assessment information may include: vehicle information, personnel location information, natural environment information and temporary operation equipment information in the section to be evaluated.

[0063] Step S300: Based on the first risk assessment value and the risk assessment information, determine the second risk assessment value corresponding to the plurality of sections to be assessed. The second risk assessment value is used to perform a fixed value assessment on the safety risk of the expressway, and the second risk assessment value can be used to more accurately assess the safety risk of the expressway and provide sufficient basis for subsequent safety risk display.

[0064] Step S400, defining a risk assessment threshold value, which is used to determine the safety risk assessment level corresponding to the second risk assessment value. By defining a risk assessment threshold value based on the safety risk situation of the expressway, the safety risk assessment level corresponding to the second risk assessment value is determined. Here, there may be multiple risk assessment threshold values, and the multiple risk assessment threshold values ​​may be arranged in order from small to large to form a risk level assessment interval. In this way, by determining the position of the second risk assessment value between the risk level assessment intervals, the risk assessment level of the assessment section of the expressway can be determined.

[0065] Step S500: Generate a security risk management map based on the security risk assessment level. Figure 2 , Figure 2 This is a schematic diagram of a safety risk management map provided in an embodiment of the present application. In order to improve the visualization capability of highway safety risk warning, the safety risk levels of different risk monitoring sections of the highway are displayed through the safety risk management map, and different risk levels correspond to different colors, so that management personnel can view the safety risks of the highway more intuitively.

[0066] In summary, the embodiment of the present application provides a highway safety warning method based on machine vision. By performing segment evaluation on multiple sections to be evaluated, the accuracy of highway safety warning can be effectively improved. The segment division can be performed based on the recognition efficiency of the machine vision algorithm, thereby ensuring the recognition efficiency of the algorithm for each section information after the segment division. Through the first risk evaluation value, various risk information in the section to be evaluated can be classified and initially evaluated, which is convenient for relevant managers to control the vehicle safety situation of the highway. The second risk evaluation value generated based on the first risk evaluation value is used to perform a fixed value evaluation on the safety risk of the highway. The second risk evaluation value can more accurately evaluate the safety risk situation of the highway and provide sufficient basis for the subsequent safety risk display. After the risk evaluation threshold is delineated, the risk evaluation level of the evaluation section of the highway is determined by determining the position of the second risk evaluation value between multiple risk evaluation thresholds. Finally, the safety risk level of different risk monitoring sections of the highway can be displayed through the safety risk management map, and different risk levels correspond to different colors, so that the management and control personnel can view the safety risk of the highway more intuitively.

[0067] In the examples of this application, please refer to Figure 3 , Figure 3 This is a flowchart of a method for obtaining risk assessment information of multiple sections to be evaluated provided by an embodiment of the present application. Step S201, obtaining vehicle information, the vehicle information including: vehicle speed information, vehicle quantity information, and vehicle congestion information; obtaining vehicle driving photos in the highway monitoring section through monitoring equipment deployed in the highway monitoring section, identifying vehicle quantity information and vehicle congestion information based on machine vision algorithms, and quantifying and outputting them;

[0068] The global positioning system (GPS) deployed in the monitoring section of the expressway monitors the vehicle position information in the monitoring section of the expressway in real time, and determines the vehicle speed information through the relevant parameters of the vehicle position information change and time change;

[0069] Step S202: obtaining personnel location information, identifying personnel features based on machine vision algorithms through monitoring equipment deployed in the monitoring section of the expressway, and obtaining personnel information in the monitoring section of the expressway;

[0070] Step S203: Acquire natural environment information, and acquire temperature information, humidity information, wind speed information and precipitation information of the section to be evaluated through the cooperation of a weather station, a temperature sensor, a humidity sensor and a wind speed sensor.

[0071] Step S204: obtain temporary operation equipment information, and monitor the operation information of the temporary operation equipment through the Internet of Things technology.

[0072] Here, when temporary operation equipment on the expressway is in operation, it should be reported in the system in accordance with the safety management requirements of the expressway. In this way, the operation of temporary equipment on the expressway can be monitored by only monitoring the temporary operation equipment information in the system.

[0073] In this way, through the coordination of multiple information monitoring means and information monitoring equipment, the highway safety early warning method can effectively identify the risk assessment information of multiple sections to be evaluated.

[0074] In this application, the vehicle quantity information and vehicle congestion information are identified based on machine vision algorithms, and the quantified output includes:

[0075] Image acquisition, using surveillance cameras and drone equipment to collect real-time image data within the monitoring section of the expressway;

[0076] Image preprocessing, adjusting the brightness and contrast of the image data, and denoising the image data. This improves image quality and enhances the accuracy of machine vision recognition. Here, the image can be scaled, cropped, and other operations can be performed according to actual needs to adapt it to the needs of algorithm processing.

[0077] Vehicle detection, applying target detection algorithms, identifying vehicles in images, locating the position of vehicles in images, and recording the position information of vehicles. For example, the YOLO algorithm can be used to identify vehicles in images, and determine the position information of vehicles on the actual road by matching the bounding box coordinates of the image with the actual coordinates.

[0078] Vehicle quantity statistics: count the vehicles in the image and obtain vehicle quantity information.

[0079] Vehicle congestion analysis, which evaluates vehicle congestion based on the distance between vehicles and lane occupancy in the image;

[0080] Vehicle congestion analysis, which evaluates vehicle congestion based on the distance between vehicles and lane occupancy;

[0081] Congestion level definition, setting congestion evaluation thresholds, and determining different congestion levels corresponding to different vehicle distances and lane occupancy rates. The congestion levels include: loose, general, congested, and very congested.

[0082] Quantitative output, converting vehicle quantity information and congestion information into specific quantitative indicator output.

[0083] It should be noted that, according to the risk assessment information, determining the first risk assessment value corresponding to the multiple sections to be evaluated includes: constructing a first risk assessment indicator system corresponding to the first risk assessment value through a cluster analysis method, and the first risk assessment system includes vehicle flow, accident rate and environmental factors. In this way, to determine the first risk assessment value, it is necessary to determine the correspondence between the risk assessment information and the vehicle flow, accident rate and environmental factors. For example, the risk assessment information includes vehicle information, personnel location information, natural environment information and temporary operation equipment information. The first risk assessment value table can be found in the table below.

[0084]

[0085]

[0086] The vehicle flow, accident rate, and environmental factors each correspond to a first risk assessment value, where the calculation formula is:

[0087] The first risk assessment value of vehicle flow = vehicle information * vehicle flow corresponding proportion 1 + personnel positioning information * vehicle flow corresponding proportion 2 + natural environment information * vehicle flow corresponding proportion 3 + temporary operation equipment information * vehicle flow corresponding proportion 4.

[0088] The first risk assessment value of the accident rate = vehicle information * accident rate corresponding proportion 1 + personnel positioning information * accident rate corresponding proportion 2 + natural environment information * accident rate corresponding proportion 3 + temporary operating equipment information * accident rate corresponding proportion 4.

[0089] The first risk assessment value of environmental factors = vehicle information * environmental factor corresponding weight 1 + personnel location information * environmental factor corresponding weight 2 + natural environment information * environmental factor corresponding weight 3 + temporary operation equipment information * environmental factor corresponding weight 4.

[0090] Here, vehicle flow corresponding weight 1 is the corresponding weight of vehicle information and vehicle flow in the first risk assessment system, vehicle flow corresponding weight 2 is the corresponding weight of personnel location information and vehicle flow in the first risk assessment system, vehicle flow corresponding weight 3 is the corresponding weight of natural environment information and vehicle flow in the first risk assessment system, and vehicle flow corresponding weight 4 is the corresponding weight of temporary operation equipment information and vehicle flow in the first risk assessment system;

[0091] The corresponding weight 1 of the accident rate is the corresponding weight of the vehicle information in the first risk assessment system and the accident rate, the corresponding weight 2 of the accident rate is the corresponding weight of the personnel positioning information in the first risk assessment system and the accident rate, the corresponding weight 3 of the accident rate is the corresponding weight of the natural environment information in the first risk assessment system and the accident rate, and the corresponding weight 4 of the accident rate is the corresponding weight of the temporary operation equipment information in the first risk assessment system and the accident rate;

[0092] The corresponding proportion 1 of environmental factors is the corresponding proportion of vehicle information and environmental factors in the first risk assessment system, the corresponding proportion 2 of environmental factors is the corresponding proportion of personnel positioning information and environmental factors in the first risk assessment system, the corresponding proportion 3 of environmental factors is the corresponding proportion of natural environment information and environmental factors in the first risk assessment system, and the corresponding proportion 4 of environmental factors is the corresponding proportion of temporary operating equipment information and environmental factors in the first risk assessment system.

[0093] In the present application, based on the first risk assessment value and risk assessment information, determining the second risk assessment value corresponding to the multiple sections to be assessed includes: based on environmental factors, determining the weight change table of the first risk assessment index system under different environments, and weighting and summing the vehicle flow, accident rate and environmental factors to determine the second risk assessment value. In this way, the second risk assessment value can be objectively and accurately determined through the weight ratio of vehicle flow, accident rate and environmental factors, thereby ensuring the application of risk rating in subsequent highway safety warnings.

[0094] In order to objectively determine the risk assessment threshold for rating the second risk assessment value, the risk assessment threshold may be determined by using a comprehensive evaluation scheme of expert evaluation and database scheme. For example, the risk assessment threshold may include:

[0095] Step S401, collecting and analyzing relevant historical data, including accident reports, traffic flow, and environmental conditions;

[0096] Step S402: Divide data features through historical data analysis to determine the initial setting of thresholds;

[0097] Step S403: verifying and adjusting the threshold through expert review and field testing;

[0098] Step S404: Set a threshold adjustment mechanism and record the threshold setting process and reasons. In this way, the risk assessment threshold can be adjusted accordingly with time and environmental changes to avoid changes in the external environment, which causes the risk threshold to be a fixed value, resulting in a distorted security risk assessment. In addition, by recording the threshold setting process and reasons, the traceability and transparency of the threshold information adjustment can be guaranteed, forming a historical data change database, and providing a more sufficient basis for the continuous adjustment of the subsequent risk threshold.

[0099] In the embodiment of the present application, there are many possible implementation methods for implementing the security risk management map. The present application only schematically illustrates one of the following methods. Generating the security risk management map may include:

[0100] Step S501: Combine the second risk assessment value with the GIS platform and use the spatial data processing and analysis capabilities of GIS.

[0101] Step S502: assign a risk level to each section to be evaluated in GIS, and mark it with different colors or symbols according to the different risk levels.

[0102] Step S503: Using the layer function of GIS, additional information layers such as traffic flow, weather conditions, topography, etc. are superimposed to provide a more comprehensive visual analysis.

[0103] Step S504: Use a data visualization tool to create an interactive map, for example, use Tableau to build an interactive map with coordinate interaction;

[0104] Step S505: Design visualization elements to represent the risk level. Figure 2 As shown, a risk multi-color map can be used to display different risk levels on the road.

[0105] Step S506: Implementing zooming, scrolling and click interaction functions of the map so that the user can explore more detailed information;

[0106] Step S507: Integrate real-time data streams, update vehicle flow information, accident rate information and environmental factor information, and dynamically create a safety risk management map;

[0107] Step S508: Setting the data refresh frequency to ensure that the map update efficiency can meet the security management needs;

[0108] Step S509: define a color code for each risk level as a quantitative indicator;

[0109] Step S510: Display the risk level of each section on the map in the form of quantitative indicators;

[0110] Step S511, integrating an early warning system on the map, and issuing a warning for a section with a risk greater than a limit threshold, where the limit threshold is the highest value among the risk assessment values ​​of the road being in a safe state;

[0111] Step S512: Design a sound warning signal and notify the on-site management personnel via SMS and phone call.

[0112] In this way, with the assistance of the safety risk management map, on-site managers can identify safety risk information on the highway in a timely manner with a faster response speed, and promptly control related dangerous items.

[0113] In summary, the embodiment of the present application provides a highway safety warning method based on machine vision. By performing segment evaluation on multiple sections to be evaluated, the accuracy of highway safety warning can be effectively improved. The segment division can be performed based on the recognition efficiency of the machine vision algorithm, thereby ensuring the recognition efficiency of the algorithm for each section information after the segment division. Through the first risk evaluation value, various risk information in the section to be evaluated can be classified and initially evaluated, which is convenient for relevant managers to control the vehicle safety situation of the highway. The second risk evaluation value generated based on the first risk evaluation value is used to perform a fixed value evaluation on the safety risk of the highway. The second risk evaluation value can more accurately evaluate the safety risk situation of the highway and provide sufficient basis for the subsequent safety risk display. After the risk evaluation threshold is delineated, the risk evaluation level of the evaluation section of the highway is determined by determining the position of the second risk evaluation value between multiple risk evaluation thresholds. Finally, the safety risk level of different risk monitoring sections of the highway can be displayed through the safety risk management map, and different risk levels correspond to different colors, so that the management and control personnel can view the safety risk of the highway more intuitively.

[0114] The present application also provides a highway safety warning device based on machine vision, including a processor, a memory, and a computer program / instruction stored in the memory, the processor is used to execute the computer program / instruction, and when the computer program / instruction is executed, the management system implements the steps of the above method.

[0115] The present application provides a computer-readable storage medium having a computer program / instruction stored thereon, wherein the computer program / instruction implements the steps of the above method when executed by a processor.

[0116] The present application further provides a computer program product, comprising a computer program / instruction, which implements the steps of the above method when executed by a processor.

[0117] In the present application, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance. The term "plurality" refers to two or more than two, unless otherwise clearly defined.

[0118] The above description is only an optional embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A highway safety warning method based on machine vision, characterized in that: Obtain the monitoring section of the expressway along the route to be inspected, and identify and warn of safety risks in the inspection section, including: Step S100, dividing the expressway along the route into sections, and dividing the expressway along the route into a plurality of sections to be evaluated; Step S200, obtaining risk assessment information of the plurality of sections to be assessed, and determining first risk assessment values ​​corresponding to the plurality of sections to be assessed according to the risk assessment information; Step S300: determining second risk assessment values ​​corresponding to the plurality of sections to be assessed based on the first risk assessment value and the risk assessment information; Step S400: defining a risk assessment threshold value, which is used to determine a security risk assessment level corresponding to the second risk assessment value; Step S500: generating a security risk management map according to the security risk assessment level; The risk assessment information includes: vehicle information, personnel location information, natural environment information and temporary operating equipment information within the section to be assessed.

2. The highway safety warning method based on machine vision according to claim 1 is characterized in that: The step of obtaining risk assessment information of multiple sections to be assessed includes: Step S201, obtaining vehicle information, the vehicle information including: vehicle speed information, vehicle quantity information, and vehicle congestion information; By means of monitoring equipment arranged in the monitoring section of the expressway, photos of vehicles traveling in the monitoring section of the expressway are obtained, and the vehicle quantity information and the vehicle congestion information are identified based on a machine vision algorithm, and the information is quantified and outputted; The vehicle position information in the monitoring section of the expressway is monitored in real time by a global positioning system (GPS) deployed in the monitoring section of the expressway, and the vehicle speed information is determined by the parameters related to the change of the vehicle position information and the change of time; Step S202: obtaining personnel location information, identifying personnel features based on machine vision algorithms through monitoring equipment deployed in the monitoring section of the expressway, and obtaining personnel information in the monitoring section of the expressway; Step S203, obtaining natural environment information, and obtaining temperature information, humidity information, wind speed information and precipitation information of the section to be evaluated through the cooperation of a weather station, a temperature sensor, a humidity sensor and a wind speed sensor; Step S204: obtain temporary operation equipment information, and monitor the operation information of the temporary operation equipment through the Internet of Things technology.

3. The highway safety warning method based on machine vision according to claim 2 is characterized in that: The identifying the vehicle quantity information and the vehicle congestion information based on a machine vision algorithm and quantifying the output includes: Image acquisition, using surveillance cameras and drone equipment to collect real-time image data within the monitoring section of the expressway; Image preprocessing, adjusting the brightness and contrast of the image data, and performing denoising on the image data; Vehicle detection, applying a target detection algorithm to identify a vehicle in an image, locate the position of the vehicle in the image, and record the position information of the vehicle; Vehicle tracking, using the optical flow method to track vehicles in consecutive frames and identify vehicles with the same features; Vehicle quantity counting, counting the vehicles in the image to obtain vehicle quantity information; Vehicle congestion analysis, evaluating vehicle congestion based on the distance between vehicles and lane occupancy in the image; Congestion level definition, setting congestion evaluation thresholds, and determining different congestion levels corresponding to different vehicle-to-vehicle distances and lane occupancy rates. The congestion levels include: loose, normal, congested, and very congested; Quantitative output, converting the vehicle quantity information and the congestion information into specific quantitative index output.

4. The highway safety warning method based on machine vision according to claim 3 is characterized in that: Determining, according to the risk assessment information, first risk assessment values ​​corresponding to the plurality of sections to be assessed comprises: constructing a first risk assessment indicator system corresponding to the first risk assessment value by a cluster analysis method, wherein the first risk assessment system comprises vehicle flow, accident rate and environmental factors; Determine the corresponding relationship between the risk assessment information and the vehicle flow, the accident rate, and the environmental factors.

5. The highway safety warning method based on machine vision according to claim 4 is characterized in that: Based on the first risk assessment value and the risk assessment information, determining the second risk assessment value corresponding to the multiple sections to be evaluated includes: based on the environmental factors, determining the weight change table of the first risk assessment index system under different environments, weightedly summing the vehicle flow, the accident rate and the environmental factors, and determining the second risk assessment value.

6. The highway safety warning method based on machine vision according to claim 5 is characterized in that: Risk assessment thresholds include: Step S401, collecting and analyzing relevant historical data, including accident reports, traffic flow, and environmental conditions; Step S402: Divide data features through historical data analysis and determine the initial setting of thresholds; Step S403: verifying and adjusting the threshold through expert review and field testing; Step S404: Set a threshold adjustment mechanism and record the threshold setting process and reasons.

7. The highway safety warning method based on machine vision according to claim 6 is characterized in that: Generating a security risk management map according to the security risk assessment level includes: Step S501, combining the second risk assessment value with the GIS platform, and utilizing the spatial data processing and analysis capabilities of GIS; Step S502: assign a risk level to each section to be evaluated in the GIS, and mark it with different colors or symbols according to the safety risk evaluation level; Step S503: Using the layer function of GIS, additional information layers such as traffic flow, weather conditions, topography, etc. are superimposed to provide a more comprehensive visual analysis; Step S504: using a data visualization tool to create an interactive map; Step S505: designing visualization elements to represent risk levels; Step S506: Implementing zooming, scrolling and click interaction functions of the map so that the user can explore more detailed information; Step S507: Integrate real-time data streams, update vehicle flow information, accident rate information and environmental factor information, and dynamically generate the safety risk management map; Step S508, setting data refresh frequency; Step S509: defining a color code as a quantitative indicator for each risk level in the security risk assessment level; Step S510: Display the risk level of each section on the map in the form of quantitative indicators; Step S511, integrating an early warning system on the map, and issuing a warning for a section with a risk greater than a limit threshold, wherein the limit threshold is the highest value among the risk assessment values ​​of the road being in a safe state; Step S512: Design a sound warning signal and notify the on-site management personnel via SMS and phone call.

8. A highway safety warning device based on machine vision, comprising a processor, a memory, and a computer program / instruction stored in the memory, characterized in that: The processor is used to execute the computer program / instructions. When the computer program / instructions are executed, the management system implements the steps of the method according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program / instruction stored thereon, characterized in that: When the computer program / instructions are executed by a processor, the steps of the method as claimed in any one of claims 1 to 7 are implemented.

10. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images