Dynamic mode recognition method and system based on deep learning

Abstract

The invention is suitable for the technical field of traffic recognition, and provides a dynamic mode recognition method and system based on deep learning, and the method comprises the steps: obtaining data information, and enabling sensors to comprise a geomagnetic sensor, a radar and a laser scanner; acquiring traffic flow information according to the vehicle information, identifying whether an abnormal condition occurs according to the traffic flow information, and adjusting traffic lights according to the abnormal condition; acquiring real-time data of the sensor, performing dynamic identification through the real-time data of the sensor, acquiring an abnormal traffic behavior according to an identification result, and sending warning information to the vehicle terminal; the method comprises the steps of obtaining traffic flow information and abnormal traffic behavior positioning data, obtaining surrounding road network information, re-planning a path based on deep learning according to the positioning data and the road network information, and sending a new path to a vehicle terminal. And the problem that the route planning cannot be updated in time after the traffic abnormal condition occurs is solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of traffic identification, and in particular relates to a dynamic pattern recognition method and system based on deep learning. Background Art

[0002] With the acceleration of urbanization and the continuous increase in the number of cars, traffic congestion and traffic safety have become important challenges faced by many cities around the world. The surge in traffic flow, especially during peak hours, has led to a series of problems such as frequent traffic accidents, road congestion, and environmental pollution. Therefore, how to achieve efficient management of road traffic and improve the safety and comfort of vehicle driving has become a hot topic in the current research on traffic management and intelligent transportation systems.

[0003] By utilizing modern information technology, communication technology, sensor technology and artificial intelligence algorithms, traffic flow, road conditions and traffic events are monitored, analyzed and dispatched in real time, thereby optimizing road traffic efficiency and reducing the incidence of traffic accidents. The core goal of the intelligent transportation system is to provide optimal traffic signal control, route planning and traffic safety management through accurate road condition prediction and dynamic adjustment.

[0004] However, when problems easily occur in the existing traffic process, it is impossible to guide the traffic to run normally in time, and the route planning cannot be updated in time after abnormal traffic conditions occur. Summary of the invention

[0005] The purpose of the embodiment of the present invention is to provide a dynamic pattern recognition method based on deep learning, aiming to solve the problems raised in the third part of the background technology.

[0006] The embodiment of the present invention is implemented as follows: a dynamic pattern recognition method based on deep learning, the method comprising:

[0007] Acquiring data information, wherein the data information includes sensor information, vehicle information and environmental information, wherein the sensor includes a geomagnetic sensor, a radar and a laser scanner;

[0008] Obtaining traffic flow information based on vehicle information, the traffic flow information being real-time traffic flow change information on the road, identifying whether abnormal conditions will occur based on the traffic flow information, the abnormal conditions including congestion, traffic accidents and traffic jam risks, and adjusting traffic lights based on the abnormal conditions;

[0009] Acquire real-time data from sensors, perform dynamic identification based on the real-time data from sensors, obtain abnormal traffic behaviors based on the identification results, the abnormal traffic behaviors include speeding, driving against traffic and fatigue driving, and send warning information to the vehicle terminal;

[0010] Obtain traffic flow information and abnormal traffic behavior location data, obtain surrounding road network information, re-plan the route based on the location data and road network information through deep learning, and send the new route to the vehicle terminal.

[0011] Preferably, the steps of obtaining traffic flow information according to vehicle information, where the traffic flow information is real-time traffic flow change information on the road, identifying whether abnormal conditions will occur according to the traffic flow information, and the abnormal conditions include congestion, car accidents and traffic jam risks, and adjusting the traffic signal lights according to the abnormal conditions specifically include:

[0012] Obtain traffic flow information according to vehicle information, and obtain the abnormal condition threshold, where the abnormal condition threshold is the critical point for the occurrence of abnormal conditions;

[0013] Compare the traffic flow information with the abnormal condition threshold to obtain the comparison result, and identify whether abnormal conditions will occur according to the comparison result;

[0014] Obtain the abnormal condition, obtain the signal light adjustment strategy, and adjust the traffic signal lights according to the abnormal condition and the adjustment strategy.

[0015] Preferably, the steps of obtaining the real-time data of the sensor, dynamically identifying through the real-time data of the sensor, obtaining abnormal traffic behaviors according to the identification result, where the abnormal traffic behaviors include speeding, reverse driving and fatigue driving, and sending a warning message to the vehicle terminal specifically include:

[0016] Obtain the real-time data of the sensor, obtain the abnormal traffic behavior threshold, and dynamically identify by comparing the real-time data of the sensor with the abnormal traffic behavior threshold;

[0017] Obtain the identification result, obtain the abnormal traffic behavior according to the identification result, and judge the specific type of the abnormal traffic behavior;

[0018] Obtain vehicle information, and send a warning message to the vehicle terminal according to the vehicle information.

[0019] Preferably, the steps of obtaining traffic flow information and abnormal traffic behavior location data, obtaining surrounding road network information, re-planning the route based on the location data and road network information through deep learning, and sending the new route to the vehicle terminal specifically include:

[0020] Obtain traffic flow information and abnormal traffic behavior location data, obtain traffic parameters according to the location data, where the traffic parameters include the number of vehicles, queue length and vehicle speed, and determine the affected area through the traffic parameters;

[0021] Obtain surrounding road network information, determine the real-time traffic flow and traffic capacity of surrounding sections through the road network information, and predict the possible spread range of congestion;

[0022] Obtain the vehicle's target location and end - point location information, re - plan the route based on deep learning according to the positioning data and road network information, and send the new route to the vehicle terminal.

[0023] Preferably, the abnormal traffic behaviors further include long - term stagnation and illegal parking.

[0024] Another object of the embodiments of the present invention is to provide a dynamic pattern recognition system based on deep learning, and the system includes:

[0025] A basic module that obtains data information, where the data information includes sensor information, vehicle information, and environmental information, and the sensors include geomagnetic sensors, radars, and lidar scanners;

[0026] A traffic flow module that obtains traffic flow information according to vehicle information. The traffic flow information is real - time traffic flow change information on the road, identifies whether abnormal conditions will occur according to the traffic flow information, and the abnormal conditions include congestion, car accidents, and traffic jam risks, and adjusts traffic signals according to the abnormal conditions;

[0027] An abnormal traffic behavior module that obtains real - time sensor data, performs dynamic recognition through the real - time sensor data, obtains abnormal traffic behaviors according to the recognition results, and the abnormal traffic behaviors include speeding, reverse driving, and fatigue driving, and sends a warning message to the vehicle terminal;

[0028] A route planning module that obtains traffic flow information and abnormal traffic behavior positioning data, obtains the surrounding road network information, re - plans the route based on deep learning according to the positioning data and road network information, and sends the new route to the vehicle terminal.

[0029] Preferably, the traffic flow module includes:

[0030] A traffic flow unit that obtains traffic flow information according to vehicle information and obtains an abnormal condition threshold, where the abnormal condition threshold is the critical point for abnormal conditions to occur;

[0031] A first comparison unit that compares the traffic flow information with the abnormal condition threshold, obtains a comparison result, and identifies whether abnormal conditions will occur according to the comparison result;

[0032] An abnormal condition unit that obtains abnormal conditions, obtains a signal light adjustment strategy, and adjusts traffic signals according to the abnormal conditions and the adjustment strategy.

[0033] Preferably, the abnormal traffic behavior module includes:

[0034] An abnormal traffic behavior unit that obtains real - time sensor data, obtains an abnormal traffic behavior threshold, and performs dynamic recognition by comparing the real - time sensor data with the abnormal traffic behavior threshold;

[0035] The second comparison unit obtains the recognition result, obtains the abnormal traffic behavior according to the recognition result, and determines the specific type of the abnormal traffic behavior;

[0036] The warning unit obtains vehicle information and sends a warning message to the vehicle terminal according to the vehicle information.

[0037] Preferably, the path planning module includes:

[0038] The positioning unit obtains traffic flow information and abnormal traffic behavior positioning data, obtains traffic parameters according to the positioning data, the traffic parameters include the number of vehicles, queue length and vehicle speed, and determines the affected area through the traffic parameters;

[0039] The road network unit obtains the surrounding road network information, determines the real-time traffic flow and traffic capacity of the surrounding road sections through the road network information, and predicts the possible spread range of congestion;

[0040] The path planning unit obtains the vehicle target position and the end position information, re-plans the path based on deep learning according to the positioning data and the road network information, and sends the new path to the vehicle terminal.

[0041] Preferably, the abnormal traffic behavior further includes long-term stagnation and illegal parking.

[0042] A dynamic pattern recognition method based on deep learning provided by an embodiment of the present invention obtains data information, the data information includes sensor information, vehicle information and environmental information, obtains traffic flow information according to the vehicle information, obtains an abnormal condition threshold, the abnormal condition threshold is the critical point of an abnormal condition, compares the traffic flow information with the abnormal condition threshold, obtains a comparison result, determines whether an abnormal condition will occur according to the comparison result, obtains the abnormal condition, obtains a traffic signal adjustment strategy, adjusts the traffic signal according to the abnormal condition and the adjustment strategy, obtains the real-time sensor data, obtains the abnormal traffic behavior threshold, performs dynamic recognition by comparing the real-time sensor data with the abnormal traffic behavior threshold, obtains the recognition result, obtains the abnormal traffic behavior according to the recognition result, determines the specific type of the abnormal traffic behavior, obtains the vehicle information, sends a warning message to the vehicle terminal according to the vehicle information, obtains the traffic flow information and the abnormal traffic behavior positioning data, obtains the traffic parameters according to the positioning data, determines the affected area through the traffic parameters, obtains the surrounding road network information, determines the real-time traffic flow and traffic capacity of the surrounding road sections through the road network information, predicts the possible spread range of congestion, obtains the vehicle target position and the end position information, re-plans the path based on deep learning according to the positioning data and the road network information, and sends the new path to the vehicle terminal, solving the problem that in the existing traffic process, when problems occur, the traffic cannot be guided to operate normally in time, and the route planning cannot be updated in time after the traffic abnormal condition occurs. Description of the Drawings

[0043] Figure 1 Flowchart of a dynamic pattern recognition method based on deep learning provided by an embodiment of the present invention;

[0044] Figure 2 Flowchart of steps for obtaining traffic flow information according to vehicle information, identifying whether an abnormal situation will occur according to the traffic flow information, and adjusting traffic lights according to the abnormal situation provided by an embodiment of the present invention;

[0045] Figure 3 Flowchart of steps for obtaining real-time sensor data, obtaining abnormal traffic behaviors according to the recognition results, and sending warning messages to vehicle terminals provided by an embodiment of the present invention;

[0046] Figure 4 Flowchart of steps for re-planning a path based on deep learning according to positioning data and road network information and sending the new path to a vehicle terminal provided by an embodiment of the present invention;

[0047] Figure 5 Architecture diagram of a dynamic pattern recognition system based on deep learning provided by an embodiment of the present invention;

[0048] Figure 6 Architecture diagram of a traffic flow module provided by an embodiment of the present invention;

[0049] Figure 7 Architecture diagram of an abnormal traffic behavior module provided by an embodiment of the present invention;

[0050] Figure 8 Architecture diagram of a path planning module provided by an embodiment of the present invention. Detailed implementation manners

[0051] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0052] It can be understood that the terms "first", "second", etc. used in the present application can be used in this document to describe various elements, but unless otherwise specified, these elements are not limited by these terms. These terms are only used to distinguish the first element from another element. For example, without departing from the scope of the present application, the first xx script can be called the second xx script, and similarly, the second xx script can be called the first xx script.

[0053] As Figure 1 shown, a dynamic pattern recognition method based on deep learning provided by an embodiment of the present invention, the method includes:

[0054] S100, acquiring data information, wherein the data information includes sensor information, vehicle information and environmental information, and the sensor includes a geomagnetic sensor, a radar and a laser scanner.

[0055] In this step, data information is obtained. The data information includes sensor information, vehicle information and environmental information. Sensors include geomagnetic sensors, radars and laser scanners. Geomagnetic sensors sense the passage of vehicles by monitoring changes in the ground magnetic field. Geomagnetic sensors are usually buried under the road surface. When a vehicle passes by, the intensity of the magnetic field sensed by the sensor will change. Usually, the duration, intensity and other characteristics of the magnetic field change are recorded to infer the speed and position of the vehicle. Radar sensors emit electromagnetic waves, which are reflected back when they encounter obstacles. Based on the time difference and frequency changes of the reflected waves, the radar can calculate the distance and relative speed of the object; the laser scanner calculates the precise position and shape of the object by emitting a laser beam and analyzing the time of the reflected light signal. The advantage of LIDAR is that it can generate very accurate three-dimensional spatial data;

[0056] Vehicle information includes the vehicle's location, speed, driving direction and license plate recognition information. The vehicle's real-time location coordinates, including latitude, longitude, altitude and other information, are obtained through the vehicle's onboard GPS device. Environmental information refers to external factors that affect traffic conditions, such as weather, road conditions, lighting conditions, etc. This information is usually obtained through environmental sensors, weather stations, traffic cameras and other equipment.

[0057] S200, obtaining traffic flow information based on vehicle information, wherein the traffic flow information is real-time traffic flow change information on the road, identifying whether abnormal conditions will occur based on the traffic flow information, wherein the abnormal conditions include congestion, traffic accidents and traffic jam risks, and adjusting traffic lights based on the abnormal conditions.

[0058] In this step, traffic flow information is obtained based on vehicle information. Vehicle information is obtained through various sensors, such as GPS, geomagnetic sensors, vehicle-mounted sensors, and radars. These sensors can provide the location, speed, direction, and acceleration of the vehicle. Based on the obtained vehicle information, real-time traffic flow can be calculated;

[0059] Based on the traffic flow information, we can identify whether there will be abnormal conditions, including congestion, traffic accidents and traffic jam risks. According to the real-time detected traffic conditions, we can adjust the traffic lights according to the abnormal conditions. Traffic lights can be adjusted to alleviate congestion, reduce accidents or respond to traffic jam risks, optimize traffic flow, alleviate congestion and improve road traffic efficiency.

[0060] S300, acquiring real-time data from sensors, performing dynamic recognition through the real-time data from sensors, acquiring abnormal traffic behaviors according to the recognition results, wherein the abnormal traffic behaviors include speeding, driving in the wrong direction, and fatigue driving, and sending warning information to the vehicle terminal.

[0061] In this step, real-time sensor data is obtained. Multiple sensors work together to collect various real-time data. Radar sensors can measure the speed and position of the vehicle in real time. Laser scanners scan the surrounding environment by emitting laser beams to obtain detailed lane information, vehicle distance and object position. Cameras use image processing technology to monitor the driving conditions of roads and vehicles in real time, identify license plates, vehicle speeds, lane deviations and other behaviors. GPS can provide accurate location information and driving trajectories, helping the system track the location of the vehicle in real time and calculate the driving speed.

[0062] Based on the sensor data collected in real time, the system can perform dynamic analysis and identify abnormal vehicle behavior, including speeding, driving in the wrong direction, and fatigue driving. Once abnormal traffic behavior is identified, a warning message is sent to the vehicle terminal to remind it to correct the improper behavior and reduce the risk of accidents.

[0063] S400, obtain traffic flow information and abnormal traffic behavior positioning data, obtain surrounding road network information, re-plan the route based on deep learning according to the positioning data and road network information, and send the new route to the vehicle terminal.

[0064] In this step, traffic flow information and abnormal traffic behavior positioning data are obtained. After detecting real-time traffic flow information and abnormal traffic behavior, if traffic jams or abnormal traffic occur, the following vehicles can be guided to detour through intelligent route replanning and rapid information dissemination, and the location of traffic jams or abnormal areas can be obtained by obtaining the specific road sections or intersections.

[0065] Obtain surrounding road network information, which includes all relevant road information, such as road section length, intersections, traffic signs, speed limit information and road conditions. Re-plan the route based on positioning data and road network information based on deep learning. Combining historical data and real-time data, the deep learning model can identify possible congestion points or traffic accident locations, and send the new route information to the vehicle terminal in real time. The driver drives along the new route and successfully avoids congested sections, saving time and improving driving safety.

[0066] like Figure 2 As shown, as a preferred embodiment of the present invention, the traffic flow information is obtained according to the vehicle information, and the traffic flow information is the real-time traffic flow change information on the road. Whether an abnormal situation will occur is identified according to the traffic flow information, and the abnormal situation includes congestion, traffic accident and traffic jam risk. The step of adjusting the traffic light according to the abnormal situation specifically includes:

[0067] S201, acquiring traffic flow information according to vehicle information, and acquiring an abnormal condition threshold, wherein the abnormal condition threshold is a critical point at which an abnormal condition occurs.

[0068] In this step, traffic flow information is obtained based on vehicle information. The real-time position, speed, and direction of the vehicle are obtained through the in-vehicle GPS system. The geomagnetic sensor can obtain the passing time and quantity of vehicles by sensing the magnetic field change when the vehicle passes by. The radar can measure the vehicle speed, distance, and traffic flow change. Through image analysis, traffic flow, vehicle speed, and road conditions can be obtained.

[0069] Based on the data obtained by the sensors, traffic flow, vehicle speed, density, and traffic capacity can be calculated, and the abnormal condition threshold can be obtained. The abnormal condition threshold refers to when certain traffic flow indicators reach a certain critical point, the traffic management system determines that an abnormal condition has occurred, such as congestion, car accident, and traffic jam risk.

[0070] The congestion threshold is usually set through indicators such as vehicle speed, traffic density, and traffic flow. When a certain indicator exceeds the set threshold, it is considered that congestion has occurred on this section of the road. The car accident threshold mainly depends on the sudden change in the driving state of the vehicle, such as a sharp drop in vehicle speed or a sharp decrease in traffic flow, which may indicate that an accident has occurred ahead. The traffic jam risk threshold is to judge whether a certain section of the road is at risk of traffic jam. Based on traffic flow prediction and traffic density, potential traffic jam points can be identified in advance.

[0071] S202. Compare the traffic flow information with the abnormal condition threshold to obtain the comparison result, and identify whether an abnormal condition will occur according to the comparison result.

[0072] In this step, compare the traffic flow information with the abnormal condition threshold. By real-time monitoring the traffic flow information and comparing it with the preset abnormal condition threshold, it can be automatically judged whether there are potential abnormal conditions on the current road, such as congestion, car accident, and traffic jam risk. The vehicle speed threshold is set at 20 km / h, the traffic density is set at more than 200 vehicles per kilometer, and the traffic flow threshold is set at 500 vehicles per hour.

[0073] Obtain the comparison result, and identify whether an abnormal condition will occur according to the comparison result. When the vehicle speed collected by the system is lower than 20 km / h, it is marked as the traffic flow is too slow. The decrease in vehicle speed usually means an increased risk of traffic congestion. When the detected traffic density exceeds 200 vehicles per kilometer, it indicates that the vehicle aggregation degree on this section of the road is very high, and there is a potential traffic congestion risk. If the traffic flow of a certain section of the road exceeds 500 vehicles per hour, it means that the number of vehicles on this section of the road exceeds its bearable traffic flow, and traffic bottlenecks or congestion may occur.

[0074] S203. Obtain the abnormal condition, obtain the traffic signal adjustment strategy, and adjust the traffic signal according to the abnormal condition and the adjustment strategy.

[0075] In this step, abnormal situations are obtained. Based on the identified abnormal situations, the traffic management system formulates corresponding signal light adjustment strategies according to different circumstances. The goal of signal light adjustment is to reduce congestion, improve traffic mobility, and ensure safety. The signal light adjustment strategies include extending the green light cycle, short-cycle adjustment, giving priority to emergency vehicles, and time-of-day control;

[0076] When the traffic flow at a certain intersection or in a certain direction is very large and traffic congestion occurs, extend the green light time in that direction so that more vehicles can pass through and waiting time can be reduced; when the traffic flow changes drastically, adopt short-cycle control to make the signal light cycle more flexible and capable of adapting to the rapid changes in traffic flow; in case of a car accident or incident, give priority to emergency vehicles such as ambulances and fire trucks to ensure the timely handling of emergencies; by analyzing the traffic flow at different times, adjust the signal light priorities at different intersections or in different directions. For example, during the morning rush hour, the main roads can be given priority to reduce interference to secondary roads.

[0077] Once an abnormal situation is identified, the system will automatically adjust the signal lights according to the above signal light adjustment strategies. When congestion occurs, extend the green light cycle or optimize the signal cycle; when a car accident occurs, give priority to emergency vehicles and close the accident section; when the risk of traffic jam appears, adjust the signal light cycle in advance to guide vehicles to detour, etc. Through these real-time responsive signal light adjustment measures, the traffic management system can effectively relieve traffic pressure, improve road traffic efficiency, reduce traffic accidents, and ensure the safety and smoothness of urban traffic.

[0078] As Figure 3 shown, as a preferred embodiment of the present invention, the steps of obtaining real-time sensor data, dynamically identifying through real-time sensor data, obtaining abnormal traffic behaviors according to the identification results, and sending warning information to the vehicle terminal specifically include:

[0079] S301, obtain real-time sensor data, obtain the threshold of abnormal traffic behaviors, and perform dynamic identification by comparing the real-time sensor data with the threshold of abnormal traffic behaviors.

[0080] In this step, real-time sensor data is obtained. Multiple sensors work together to collect various real-time data. The radar sensor can measure the speed and position of the vehicle in real time. The laser scanner scans the surrounding environment by emitting laser beams to obtain detailed lane information, vehicle distances, and object positions. The camera can, through image processing technology, monitor the road and the driving conditions of vehicles in real time, identify behaviors such as license plate recognition, vehicle speed, and lane departure. The GPS can provide accurate position information and driving trajectories, helping the system to track the position of the vehicle in real time and calculate the driving speed;

[0081] Obtain the threshold for abnormal traffic behavior. The threshold for abnormal traffic behavior is set based on traffic safety standards and can be dynamically adjusted according to different traffic scenarios, road types, time periods, etc. For example, for the speeding threshold, on urban roads, the speed limit may be 60 km / h, and on highways, the speed limit may be 120 km / h; the reverse driving threshold is the situation where the traffic flow direction is opposite; the fatigue driving threshold can be that the eyes are closed for more than 5 seconds or yawning continuously 3 times, which means there is a risk of fatigue driving.

[0082] Dynamically identify abnormal traffic behavior by comparing the real-time data of sensors with the threshold for abnormal traffic behavior. After obtaining the real-time data of sensors and the set threshold, perform dynamic comparison to determine whether the current traffic behavior is abnormal. If the data of a certain behavior exceeds the set threshold, an alarm will be triggered or corresponding measures will be taken.

[0083] S302, Obtain the recognition result, obtain the abnormal traffic behavior according to the recognition result, and judge the specific type of the abnormal traffic behavior.

[0084] In this step, obtain the recognition result. Through the comparison and analysis of data, the recognition result can be classified into different abnormal traffic behaviors. Abnormal traffic behaviors include speeding, reverse driving, and fatigue driving. Judge the specific type of the abnormal traffic behavior according to the recognition result;

[0085] Through the real-time analysis, processing, and comparison of sensor data, traffic abnormal behaviors can be accurately identified and classified according to different judgment criteria. Each type of abnormal behavior has different characteristics and thresholds, and the system will react according to the real-time data and the set thresholds.

[0086] For example, it is monitored that the speed of a vehicle is 130 km / h (the speeding threshold is set at 100 km / h), and at the same time, it is monitored by a camera that the vehicle is reversing on a one-way road. First, it is recognized as a speeding behavior. Then, the system detects that the driving direction of the vehicle is opposite to the road lane direction and determines it as a reverse driving behavior.

[0087] S303, Obtain vehicle information, and send a warning message to the vehicle terminal according to the vehicle information.

[0088] In this step, obtain vehicle information. Through the vehicle information, the driver information of the vehicle can be known, and at the same time, the vehicle's on-board information can be obtained. Through the vehicle's on-board information, the vehicle's on-board audio system can be connected. Once the system identifies an abnormal traffic behavior or potential risk, the intelligent transportation system needs to immediately send a warning message to the driver to remind him to take measures to ensure driving safety;

[0089] For example, on a road with a speed limit of 80 km / h, the system detects the real-time speed of a vehicle as 95 km / h through radar or GPS, which exceeds the speed limit. Immediately, a warning is triggered and played through the in-vehicle audio: Warning: Your vehicle speed has exceeded 80 km / h. Please slow down.

[0090] As Figure 4 shown, as a preferred embodiment of the present invention, the steps of obtaining traffic flow information and abnormal traffic behavior location data, obtaining surrounding road network information, re-planning a route based on the location data and road network information by deep learning, and sending the new route to the vehicle terminal specifically include:

[0091] S401, obtain traffic flow information and abnormal traffic behavior location data, obtain traffic parameters according to the location data, where the traffic parameters include the number of vehicles, queue length, and vehicle speed, and determine the affected area through the traffic parameters.

[0092] In this step, after obtaining traffic flow information and abnormal traffic behavior location data, when real-time traffic flow information and abnormal traffic behavior are detected, in case of traffic jams or abnormal traffic, the rear vehicles can be guided to detour through intelligent route re-planning and rapid information dissemination. The location of the traffic jam or abnormal area can be obtained by getting the specific road section or intersection location;

[0093] Obtain traffic parameters according to the location data. The number of vehicles refers to the total number of vehicles passing through or staying at a specific location within a specific time; the queue length refers to the length of the vehicle queue caused by traffic congestion or other reasons; the vehicle speed refers to the speed at which the vehicle is traveling. When the vehicle speed is lower than the specified or average speed, it usually indicates that there may be traffic congestion or a traffic accident;

[0094] Determine the affected area through the traffic parameters. By obtaining the affected area, the scope of the impact can be known. For short-term traffic congestion caused by a certain unexpected event, such as a traffic accident and vehicle breakdown, the identification of the affected area can help the traffic management system respond quickly and take timely measures, such as adjusting traffic lights, issuing detour instructions, or increasing traffic control, etc., so as to effectively relieve congestion, improve traffic efficiency, and ensure driving safety.

[0095] S402, obtain surrounding road network information, determine the real-time traffic flow and traffic capacity of the surrounding road sections through the road network information, and predict the possible spread range of congestion.

[0096] In this step, obtain the information of the surrounding road network. The information of the surrounding road network refers to the information of all traffic infrastructure such as roads, intersections, road sections, traffic signs, speed limits, etc. involved in the traffic flow of a certain area. Determine the real-time traffic flow and traffic capacity of the surrounding road sections through the road network information, and find out whether there is a more unobstructed route for the surrounding routes, so as to accurately evaluate the current traffic capacity of the road. Compare the remaining routes with the existing routes and compare the traffic capacities;

[0097] Predict the possible spread range of congestion. Through the road network structure, traffic flow data and real-time traffic events, it is possible to predict whether the traffic conditions of a certain road section will spread to other road sections or a larger area. For example, some road sections are connected to each other. If a road section is congested, it may quickly affect other road sections. The main roads, intersections, etc. in the traffic network are often the key areas for congestion spread; emergencies such as traffic accidents will cause local congestion and quickly spread. Especially when a road section is closed or the lanes are restricted, the traffic flow on the surrounding roads will increase sharply.

[0098] Suppose there is a traffic accident in the city center, resulting in a sharp drop in the vehicle speed at a major intersection. Through real-time monitoring data and traffic flow prediction models, the system can identify this situation and predict that the congestion will spread to the two surrounding main roads. Based on this prediction, the traffic management system can adjust the signal timing in advance, suggest that drivers choose alternative routes, and successfully avoid the spread of congestion.

[0099] S403, obtain the vehicle target location and destination location information, re-plan the path based on deep learning according to the positioning data and road network information, and send the new path to the vehicle terminal.

[0100] In this step, obtain the vehicle target location and destination location information. The current location of the vehicle is usually obtained by on-vehicle sensors. The GPS system can provide the real-time longitude and latitude information of the vehicle. The target destination location of the vehicle is usually input by the driver or the autonomous driving system, or set through the destination of the navigation system;

[0101] Re-plan the path based on deep learning according to the positioning data and road network information. Learn the traffic flow data through the neural network, predict the future traffic conditions, and select the un-congested path. When the deep learning model completes the path planning, the system will generate the optimal driving path and transmit the new path information to the vehicle terminal in real time through the on-vehicle communication system. After receiving the path information, the vehicle terminal updates the navigation display and adjusts the driving route according to the real-time traffic conditions.

[0102] As Figure 5 shown, a dynamic pattern recognition system based on deep learning provided by an embodiment of the present invention, the system includes:

[0103] The basic module 100 is used to obtain data information, and the data information includes sensor information, vehicle information and environmental information. The sensors include geomagnetic sensors, radars and lidar scanners.

[0104] In this system, the basic module 100 obtains data information, which includes sensor information, vehicle information and environmental information. The sensors include geomagnetic sensors, radars and lidar scanners. The geomagnetic sensor perceives the passing of vehicles by monitoring the changes in the ground magnetic field. Geomagnetic sensors are usually buried under the road surface. When a vehicle passes by, the magnetic field intensity sensed by the sensor will change. Usually, the duration, intensity and other characteristics of the magnetic field change are recorded, and then the speed and position of the vehicle are deduced. The radar sensor emits electromagnetic waves, and when the electromagnetic waves encounter obstacles, they will be reflected back. According to the time difference and frequency change of the reflected wave, the radar can deduce the distance and relative speed of the object; the lidar scanner calculates the precise position and shape of the object by emitting laser beams and analyzing the time of the reflected optical signals. The advantage of lidar is that it can generate very precise three-dimensional spatial data;

[0105] The vehicle information includes the position, speed, driving direction and license plate recognition information of the vehicle. The real-time position coordinates of the vehicle are obtained through in-vehicle GPS devices, including longitude, latitude, altitude and other information. The environmental information refers to external factors that affect traffic conditions, such as weather, road conditions, lighting conditions, etc. This information is usually obtained through devices such as environmental sensors, weather stations, and traffic cameras.

[0106] The traffic flow module 200 is used to obtain traffic flow information according to the vehicle information. The traffic flow information is the real-time flow change information on the road. Whether abnormal conditions will occur is identified according to the traffic flow information. The abnormal conditions include congestion, car accidents and traffic jam risks. The traffic lights are adjusted according to the abnormal conditions.

[0107] In this system, the traffic flow module 200 obtains traffic flow information according to the vehicle information. The vehicle information is obtained through multiple sensors, such as GPS, geomagnetic sensors, in-vehicle sensors, and radars. These sensors can provide the position, speed, direction and acceleration of the vehicle. Based on the obtained vehicle information, the real-time traffic flow can be calculated;

[0108] Whether abnormal conditions will occur is identified according to the traffic flow information. The abnormal conditions include congestion, car accidents and traffic jam risks. According to the real-time detected traffic conditions, the traffic lights are adjusted according to the abnormal conditions. The traffic lights can be adjusted to relieve congestion, reduce accidents or cope with traffic jam risks, optimize traffic flow, relieve congestion and improve road traffic efficiency.

[0109] The abnormal traffic behavior module 300 is used to obtain real-time sensor data, perform dynamic identification based on the real-time sensor data, obtain abnormal traffic behaviors according to the identification results, where the abnormal traffic behaviors include speeding, reverse driving, and fatigue driving, and send warning information to the vehicle terminal.

[0110] In this system, the abnormal traffic behavior module 300 obtains real-time sensor data. Multiple sensors work together to collect various real-time data. The radar sensor can measure the speed and position of the vehicle in real time. The laser scanner scans the surrounding environment by emitting laser beams to obtain detailed lane information, vehicle distance, and object positions. The camera can, through image processing technology, monitor the road and the driving conditions of the vehicle in real time, and identify behaviors such as license plate recognition, vehicle speed, and lane departure. The GPS can provide accurate position information and driving trajectories, helping the system to track the vehicle's position in real time and calculate the driving speed;

[0111] According to the real-time sensor data collected, the system can perform dynamic analysis to identify abnormal behaviors of the vehicle, including speeding, reverse driving, and fatigue driving. Once an abnormal traffic behavior is identified, a warning information is sent to the vehicle terminal to remind it to correct the improper behavior and reduce the risk of accidents.

[0112] The path planning module 400 is used to obtain traffic flow information and abnormal traffic behavior positioning data, obtain the surrounding road network information, re-plan the path based on deep learning according to the positioning data and the road network information, and send the new path to the vehicle terminal.

[0113] In this system, the path planning module 400 obtains traffic flow information and abnormal traffic behavior positioning data. After detecting the real-time traffic flow information and abnormal traffic behaviors, if there is a traffic jam or abnormal traffic, it can guide the following vehicles to detour through intelligent route re-planning and rapid information dissemination, and obtain the positioning of the traffic jam or abnormal area by getting the specific road section or intersection position;

[0114] Obtain the surrounding road network information. The surrounding road network information includes all relevant road information, such as road section length, intersections, traffic signs, speed limits, and road conditions. Based on deep learning, according to the positioning data and the road network information, re-plan the path. Combining historical data and real-time data, the deep learning model can identify possible congestion points or traffic accident locations, and send the new path information to the vehicle terminal in real time. The driver drives according to the new path, successfully avoiding the congested road section, saving time and improving driving safety.

[0115] As Figure 6 shown, as a preferred embodiment of the present invention, the traffic flow module 200 includes:

[0116] A traffic flow unit 201, which is used to obtain traffic flow information according to vehicle information and obtain an abnormal condition threshold, where the abnormal condition threshold is the critical point for the occurrence of an abnormal condition.

[0117] In this module, the traffic flow unit 201 obtains traffic flow information according to vehicle information. The real-time position, speed and direction of the vehicle are obtained through the in-vehicle GPS system. The geomagnetic sensor can obtain the passing time and quantity of vehicles by sensing the change of the magnetic field when the vehicle passes by. The radar can measure the speed, distance and traffic flow change of the vehicle. Through image analysis, the traffic flow, vehicle speed and road conditions can be obtained.

[0118] Based on the data obtained by the sensors, the traffic flow, vehicle speed, density and traffic capacity can be calculated, and the abnormal condition threshold can be obtained. The abnormal condition threshold refers to that when certain traffic flow indicators reach a certain critical point, the traffic management system determines that an abnormal condition has occurred, such as congestion, car accident and traffic jam risk.

[0119] The congestion threshold is usually set by indicators such as vehicle speed, traffic density, traffic flow, etc. When a certain indicator exceeds the set threshold, it is considered that congestion has occurred on this section of the road; the car accident threshold mainly depends on the sudden change of the vehicle driving state, such as the sharp decrease of the vehicle speed or the sharp decrease of the traffic flow, which may indicate that an accident has occurred ahead; the traffic jam risk threshold is to judge whether a certain section of the road faces the risk of traffic jam. Based on traffic flow prediction and traffic density, potential traffic jam points can be identified in advance.

[0120] A first comparison unit 202, which is used to compare the traffic flow information with the abnormal condition threshold, obtain a comparison result, and identify whether an abnormal condition will occur according to the comparison result.

[0121] In this module, the first comparison unit 202 compares the traffic flow information with the abnormal condition threshold. By real-time monitoring the traffic flow information and comparing it with the preset abnormal condition threshold, it can automatically judge whether there are potential abnormal conditions on the current road, such as congestion, car accident and traffic jam risk. The vehicle speed threshold is set to 20 km / h, the traffic density is set to more than 200 vehicles per kilometer, and the traffic flow threshold is set to 500 vehicles per hour.

[0122] Obtain the comparison result, and identify whether an abnormal condition will occur according to the comparison result. When the vehicle speed collected by the system is lower than 20 km / h, it is marked as the traffic flow is too slow. The decrease of the vehicle speed usually means an increased risk of traffic congestion; when the detected traffic density exceeds 200 vehicles per kilometer, it indicates that the vehicle aggregation degree on this section of the road is very high and there is a potential traffic congestion risk; if the traffic flow of a certain section of the road exceeds 500 vehicles per hour, it means that the number of vehicles on this section of the road exceeds its bearable traffic flow and traffic bottlenecks or congestion may occur.

[0123] Anomaly situation unit 203 is used to obtain anomaly situations, obtain traffic signal adjustment strategies, and adjust traffic signals according to the anomaly situations and adjustment strategies.

[0124] In this module, the anomaly situation unit 203 obtains anomaly situations. According to the identified anomaly situations, the traffic management system formulates corresponding traffic signal adjustment strategies according to different situations. The goal of traffic signal adjustment is to reduce congestion, improve traffic mobility, and ensure safety. The traffic signal adjustment strategies include extending the green light cycle, short-cycle adjustment, giving priority to emergency vehicles, and time-of-day control.

[0125] When the traffic flow at a certain intersection or in a certain direction is very large and traffic congestion occurs, extend the green light time in that direction so that more vehicles can pass through and reduce waiting time. When the traffic flow changes violently, adopt short-cycle control to make the signal cycle more flexible and adapt to the rapid change of traffic flow. When a car accident or incident occurs, give priority to emergency vehicles such as ambulances and fire trucks to ensure the timely handling of emergencies. By analyzing the traffic flow at different times, adjust the signal priorities at different intersections or in different directions. For example, during the morning rush hour, the main roads can be given priority to reduce interference to secondary roads.

[0126] Once an anomaly situation is identified, the system will automatically adjust the traffic signals according to the above traffic signal adjustment strategies. When congestion occurs, extend the green light cycle or optimize the signal cycle. When a car accident occurs, give priority to emergency vehicles and close the accident section. When the risk of traffic jams appears, adjust the signal cycle in advance to guide vehicles to detour, etc. Through these real-time responsive traffic signal adjustment measures, the traffic management system can effectively relieve traffic pressure, improve road traffic efficiency, reduce traffic accidents, and ensure the safety and smoothness of urban traffic.

[0127] Such as Figure 7 shown, as a preferred embodiment of the present invention, the abnormal traffic behavior module 300 includes:

[0128] An abnormal traffic behavior unit 301 is used to obtain real-time sensor data, obtain abnormal traffic behavior thresholds, and perform dynamic identification by comparing the real-time sensor data with the abnormal traffic behavior thresholds.

[0129] In this module, the abnormal traffic behavior unit 301 obtains real-time sensor data. Multiple sensors work together to collect various real-time data. The radar sensor can measure the speed and position of vehicles in real time. The laser scanner scans the surrounding environment by emitting laser beams to obtain detailed lane information, vehicle distances, and object positions. The camera can, through image processing technology, monitor the road and the driving conditions of vehicles in real time, identify behaviors such as license plates, vehicle speeds, and lane departures. The GPS can provide accurate position information and driving trajectories to help the system track the position of vehicles in real time and calculate the driving speed.

[0130] Obtain the thresholds for abnormal traffic behaviors. The thresholds for abnormal traffic behaviors are set based on traffic safety standards and can be dynamically adjusted according to different traffic scenarios, road types, time periods, etc. For example, for the speeding threshold, on urban roads, the speed limit may be 60 km / h, and on highways, the speed limit may be 120 km / h; the reverse driving threshold is the situation where the traffic flow direction is opposite; the fatigue driving threshold can be that the eyes are closed for more than 5 seconds or yawning continuously 3 times, that is, it is considered that there is a risk of fatigue driving.

[0131] Dynamically identify by comparing the real-time data of sensors with the thresholds for abnormal traffic behaviors. After obtaining the real-time data of sensors and the set thresholds, perform dynamic comparison to determine whether the current traffic behavior is abnormal. If the data of a certain behavior exceeds the set threshold, an alarm will be triggered or corresponding measures will be taken.

[0132] The second comparison unit 302 is used to obtain the recognition result, obtain the abnormal traffic behavior according to the recognition result, and judge the specific type of the abnormal traffic behavior.

[0133] In this module, the second comparison unit 302 obtains the recognition result. Through the comparison and analysis of data, the recognition result can be classified into different abnormal traffic behaviors. Abnormal traffic behaviors include speeding, reverse driving, and fatigue driving. Judge the specific type of the abnormal traffic behavior according to the recognition result;

[0134] Through the real-time analysis, processing, and comparison of sensor data, traffic abnormal behaviors can be accurately identified and classified according to different judgment criteria. Each type of abnormal behavior has different characteristics and thresholds, and the system will make responses according to the real-time data and the set thresholds.

[0135] For example, it is monitored that the speed of a vehicle is 130 km / h (the speeding threshold is set at 100 km / h), and at the same time, it is monitored through the camera that the vehicle is driving in reverse on a one-way road. First, it is recognized as a speeding behavior. Then, the system detects that the driving direction of the vehicle is opposite to the road lane direction and determines it as a reverse driving behavior.

[0136] The warning unit 303 is used to obtain vehicle information and send a warning message to the vehicle terminal according to the vehicle information.

[0137] In this module, the warning unit 303 obtains vehicle information, knows the driver information of the vehicle through the vehicle information, and at the same time obtains the vehicle-mounted information of the vehicle. Through the vehicle-mounted information of the vehicle, the vehicle-mounted audio system of the vehicle can be connected. Once the system identifies an abnormal traffic behavior or a potential risk, the intelligent transportation system needs to immediately send a warning message to the driver to remind him to take measures to ensure driving safety;

[0138] For example, on a road with a speed limit of 80 km / h, the system detects the real-time speed of a vehicle as 95 km / h through radar or GPS, which exceeds the speed limit. Immediately, a warning is triggered and played through the in-vehicle audio: Warning: Your vehicle speed has exceeded 80 km / h. Please slow down.

[0139] As Figure 8 shown, as a preferred embodiment of the present invention, the path planning module 400 includes:

[0140] A positioning unit 401, configured to obtain traffic flow information and abnormal traffic behavior positioning data, obtain traffic parameters according to the positioning data, where the traffic parameters include the number of vehicles, queue length, and vehicle speed, and determine the affected area through the traffic parameters.

[0141] In this module, the positioning unit 401 obtains traffic flow information and abnormal traffic behavior positioning data. After detecting real-time traffic flow information and abnormal traffic behavior, in case of traffic jams or abnormal traffic, the rear vehicles can be guided to detour through intelligent route re-planning and rapid information dissemination, and the location of the traffic jam or abnormal area can be obtained by acquiring the specific road section or intersection position;

[0142] Traffic parameters are obtained according to the positioning data. The number of vehicles refers to the total number of vehicles passing through or staying at a specific location within a specific time. The queue length refers to the length of the vehicle queue caused by traffic congestion or other reasons. The vehicle speed refers to the speed at which the vehicle is traveling. When the vehicle speed is lower than the specified or average speed, it usually indicates that there may be traffic congestion or traffic accidents;

[0143] The affected area is determined through the traffic parameters. By obtaining the affected area, the scope of the impact can be known. For short-term traffic congestion caused by a certain unexpected event, such as a traffic accident and vehicle breakdown, the identification of the affected area can help the traffic management system to react quickly and take timely measures, such as adjusting traffic lights, issuing detour instructions, or increasing traffic control, so as to effectively relieve congestion, improve traffic efficiency, and ensure driving safety.

[0144] A road network unit 402, configured to obtain surrounding road network information, determine the real-time traffic flow and traffic capacity of surrounding road sections through the road network information, and predict the possible spread range of congestion.

[0145] In this module, the road network unit 402 obtains surrounding road network information. The surrounding road network information refers to information on all roads, intersections, road section information, traffic signs, speed limits, and other traffic infrastructure related to traffic flow in a certain area. The real-time traffic flow and traffic capacity of surrounding road sections are determined through the road network information, and whether there is a more unobstructed route for the surrounding routes can be obtained, and the current traffic capacity of the road can be accurately evaluated, and the remaining routes are compared with the existing routes to compare the traffic capacity;

[0146] Predict the possible scope of congestion spread. Through the road network structure, traffic flow data, and real-time traffic events, it is possible to predict whether the traffic conditions on a certain road section will spread to other road sections or a larger area. For example, some road sections are interconnected. If a congestion occurs on one road section, it may quickly affect other road sections. Main roads, intersections, etc. in the traffic network are often key areas for congestion spread; emergencies such as traffic accidents will cause local congestion and quickly spread. Especially when a road section is closed or the lanes are restricted, the traffic flow on the surrounding roads will increase sharply.

[0147] Suppose a traffic accident occurs in the city center, resulting in a sharp drop in the vehicle speed at a major intersection. Through real-time monitoring data and traffic flow prediction models, the system can identify this situation and predict that the congestion will spread to the two surrounding main roads. Based on this prediction, the traffic management system can adjust the signal timing in advance, suggest alternative routes to drivers, and successfully avoid the spread of congestion.

[0148] The path planning unit 403 is used to obtain the vehicle target location and the end location information, re-plan the path based on deep learning according to the positioning data and road network information, and send the new path to the vehicle terminal.

[0149] In this module, the path planning unit 403 obtains the vehicle target location and the end location information. The current location of the vehicle is usually obtained by on-vehicle sensors. The GPS system can provide the real-time longitude and latitude information of the vehicle. The target end location of the vehicle is usually input by the driver or the autonomous driving system, or set through the destination of the navigation system;

[0150] Re-plan the path based on deep learning according to the positioning data and road network information. Learn the traffic flow data through a neural network, predict the future traffic conditions, and select an un-congested path. When the deep learning model completes the path planning, the system will generate the optimal driving path and transmit the new path information to the vehicle terminal in real time through the on-vehicle communication system. After receiving the path information, the vehicle terminal updates the navigation display and adjusts the driving route according to the real-time traffic conditions.

[0151] In one embodiment, a computer device is proposed. The computer device includes a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the following steps are implemented:

[0152] Obtain data information, where the data information includes sensor information, vehicle information, and environmental information. The sensors include geomagnetic sensors, radars, and lidar scanners;

[0153] Obtaining traffic flow information based on vehicle information, the traffic flow information being real-time traffic flow change information on the road, identifying whether abnormal conditions will occur based on the traffic flow information, the abnormal conditions including congestion, traffic accidents and traffic jam risks, and adjusting traffic lights based on the abnormal conditions;

[0154] Acquire real-time data from sensors, perform dynamic identification based on the real-time data from sensors, obtain abnormal traffic behaviors based on the identification results, the abnormal traffic behaviors include speeding, driving against traffic and fatigue driving, and send warning information to the vehicle terminal;

[0155] Obtain traffic flow information and abnormal traffic behavior positioning data, obtain surrounding road network information, re-plan the route based on positioning data and road network information based on deep learning, and send the new route to the vehicle terminal.

[0156] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the processor performs the following steps:

[0157] Acquiring data information, wherein the data information includes sensor information, vehicle information and environmental information, wherein the sensor includes a geomagnetic sensor, a radar and a laser scanner;

[0158] Obtaining traffic flow information based on vehicle information, the traffic flow information being real-time traffic flow change information on the road, identifying whether abnormal conditions will occur based on the traffic flow information, the abnormal conditions including congestion, traffic accidents and traffic jam risks, and adjusting traffic lights based on the abnormal conditions;

[0159] Acquire real-time data from sensors, perform dynamic identification based on the real-time data from sensors, obtain abnormal traffic behaviors based on the identification results, the abnormal traffic behaviors include speeding, driving against traffic and fatigue driving, and send warning information to the vehicle terminal;

[0160] Obtain traffic flow information and abnormal traffic behavior positioning data, obtain surrounding road network information, re-plan the route based on positioning data and road network information based on deep learning, and send the new route to the vehicle terminal.

[0161] It should be understood that although the steps in the flowcharts of the embodiments of the present invention are shown in sequence according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless there is a clear description in this article, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, at least a part of the steps in each embodiment may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed alternately or in turn with at least a part of other steps or sub-steps or stages of other steps.

[0162] Those of ordinary skill in the art can understand that all or part of the processes of implementing the methods in the above embodiments can be completed by instructing relevant hardware through a computer program. The program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, storage, database, or other medium used in the embodiments provided in the present application can include non-volatile and / or volatile memories. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0163] The technical features of the above-described embodiments can be combined arbitrarily. For the sake of concise description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered to be within the scope described in this specification.

[0164] The above-described embodiments merely represent several implementation manners of the present invention. The description thereof is relatively specific and detailed, but it should not be construed as a limitation to the scope of the patent of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all fall within the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the appended claims.

[0165] The above is only the preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A dynamic pattern recognition method based on deep learning, characterized in that: The method comprises: Acquiring data information, wherein the data information includes sensor information, vehicle information and environmental information, wherein the sensor includes a geomagnetic sensor, a radar and a laser scanner; Obtaining traffic flow information based on vehicle information, the traffic flow information being real-time traffic flow change information on the road, identifying whether abnormal conditions will occur based on the traffic flow information, the abnormal conditions including congestion, traffic accidents and traffic jam risks, and adjusting traffic lights based on the abnormal conditions; Acquire real-time data from sensors, perform dynamic identification based on the real-time data from sensors, obtain abnormal traffic behaviors based on the identification results, the abnormal traffic behaviors include speeding, driving against traffic and fatigue driving, and send warning information to the vehicle terminal; Obtain traffic flow information and abnormal traffic behavior positioning data, obtain surrounding road network information, re-plan the route based on positioning data and road network information based on deep learning, and send the new route to the vehicle terminal.

2. The dynamic pattern recognition method based on deep learning according to claim 1, characterized in that: The step of obtaining traffic flow information according to vehicle information, wherein the traffic flow information is real-time traffic flow change information on the road, identifying whether an abnormal situation will occur according to the traffic flow information, wherein the abnormal situation includes congestion, traffic accident and traffic jam risk, and adjusting the traffic light according to the abnormal situation specifically includes: Acquire traffic flow information according to vehicle information, and acquire an abnormal condition threshold, wherein the abnormal condition threshold is a critical point at which an abnormal condition occurs; Compare the traffic flow information with the abnormal condition threshold, obtain the comparison result, and identify whether an abnormal condition will occur based on the comparison result; Obtain abnormal conditions, obtain traffic light adjustment strategies, and adjust traffic lights according to the abnormal conditions and adjustment strategies.

3. The dynamic pattern recognition method based on deep learning according to claim 1, characterized in that: The step of acquiring real-time sensor data, performing dynamic identification through the real-time sensor data, acquiring abnormal traffic behavior according to the identification result, wherein the abnormal traffic behavior includes speeding, driving against traffic and fatigue driving, and sending warning information to the vehicle terminal specifically includes: Obtain real-time sensor data, obtain abnormal traffic behavior thresholds, and dynamically identify abnormal traffic behavior thresholds by comparing real-time sensor data with the abnormal traffic behavior thresholds; Obtain recognition results, obtain abnormal traffic behavior based on the recognition results, and determine the specific type of abnormal traffic behavior; Obtain vehicle information and send warning information to the vehicle terminal based on the vehicle information.

4. The dynamic pattern recognition method based on deep learning according to claim 1, characterized in that: The steps of obtaining traffic flow information and abnormal traffic behavior positioning data, obtaining surrounding road network information, replanning the route based on the positioning data and road network information based on deep learning, and sending the new route to the vehicle terminal specifically include: Obtaining traffic flow information and abnormal traffic behavior location data, obtaining traffic parameters based on the location data, the traffic parameters including the number of vehicles, queue length and vehicle speed, and determining the affected area through the traffic parameters; Obtain surrounding road network information, determine the real-time traffic flow and capacity of surrounding roads through the road network information, and predict the possible spread of congestion; Obtain the vehicle's target location and terminal location information, re-plan the route based on positioning data and road network information based on deep learning, and send the new route to the vehicle terminal.

5. The dynamic pattern recognition method based on deep learning according to claim 1, characterized in that: The abnormal traffic behavior also includes long-term stagnation and illegal parking.

6. A dynamic pattern recognition system based on deep learning, characterized in that: The system comprises: A basic module, which acquires data information, wherein the data information includes sensor information, vehicle information and environmental information, and the sensors include geomagnetic sensors, radars and laser scanners; Traffic flow module, which obtains traffic flow information based on vehicle information. The traffic flow information is real-time traffic flow change information on the road. It identifies whether abnormal conditions will occur based on the traffic flow information. The abnormal conditions include congestion, traffic accidents and traffic jam risks, and adjusts traffic lights according to the abnormal conditions. The abnormal traffic behavior module obtains real-time data from sensors, performs dynamic identification based on the real-time data from sensors, obtains abnormal traffic behaviors based on the identification results, and the abnormal traffic behaviors include speeding, driving against traffic and fatigue driving, and sends warning information to the vehicle terminal; The path planning module obtains traffic flow information and abnormal traffic behavior positioning data, obtains surrounding road network information, replans the path based on the positioning data and road network information based on deep learning, and sends the new path to the vehicle terminal.

7. A deep learning-based dynamic pattern recognition system according to claim 6, characterized in that: The traffic flow module includes: A traffic flow unit, which obtains traffic flow information according to vehicle information and obtains an abnormal condition threshold, wherein the abnormal condition threshold is a critical point at which an abnormal condition occurs; A first comparison unit compares the traffic flow information with an abnormal condition threshold, obtains a comparison result, and identifies whether an abnormal condition will occur according to the comparison result; The abnormal condition unit obtains the abnormal condition, obtains the traffic light adjustment strategy, and adjusts the traffic light according to the abnormal condition and the adjustment strategy.

8. A deep learning-based dynamic pattern recognition system according to claim 7, characterized in that: The abnormal traffic behavior module includes: The abnormal traffic behavior unit obtains real-time data from sensors, obtains abnormal traffic behavior thresholds, and dynamically identifies abnormal traffic behavior thresholds by comparing real-time data from sensors; The second comparison unit obtains the recognition result, obtains the abnormal traffic behavior according to the recognition result, and determines the specific type of the abnormal traffic behavior; The warning unit obtains vehicle information and sends warning information to the vehicle terminal according to the vehicle information.

9. A deep learning-based dynamic pattern recognition system according to claim 8, characterized in that: The path planning module includes: A positioning unit, which obtains traffic flow information and abnormal traffic behavior positioning data, obtains traffic parameters according to the positioning data, and the traffic parameters include the number of vehicles, queue length and vehicle speed, and determines the affected area through the traffic parameters; The road network unit obtains the surrounding road network information, determines the real-time traffic flow and capacity of the surrounding road sections through the road network information, and predicts the possible spread of congestion; The path planning unit obtains the vehicle's target location and terminal location information, replans the path based on positioning data and road network information based on deep learning, and sends the new path to the vehicle terminal.

10. A deep learning-based dynamic pattern recognition system according to claim 9, characterized in that: The abnormal traffic behavior also includes long-term stagnation and illegal parking.

Images