Intelligent traffic monitoring method and system based on Internet of Things

Abstract

The present invention discloses an intelligent traffic monitoring method and system based on the Internet of Things. The intelligent traffic monitoring method comprises the following steps: step one: distinguishing non-motor vehicles by combining the angle between a seat and a human body and a driving speed; step two: detecting the illegal behavior of non-motor vehicles and issuing warnings for the illegal behavior; step three: recording the illegal information of non-motor vehicles and generating corresponding feature information; step four: performing illegal information statistics and action trajectory analysis on vehicles in the archives, wherein the non-motor vehicle classification and distinction module is used to capture non-motor vehicle images in real time and accurately classify electric bicycles and bicycles through advanced image processing technology, and the illegal behavior detection and warning module is used to accurately detect the illegal behavior of non-motor vehicles through traffic light status detection and face / license plate recognition. The present invention has the characteristics of improving the accuracy and real-time performance of illegal detection.

Description

Technical Field

[0001] The present invention relates to the technical field of traffic monitoring, and in particular to an intelligent traffic monitoring method and system based on the Internet of Things. Background Art

[0002] As urban traffic becomes increasingly busy, the problem of non-motor vehicle violations has gradually become prominent, and traditional monitoring methods are difficult to meet the needs of efficient management. At present, the supervision of traffic violations in some cities mainly relies on manual patrols and limited surveillance cameras. This method has problems such as blind spots in monitoring, low efficiency, and high dependence on human resources, which makes it difficult to effectively maintain traffic order. In order to solve these problems, some traffic violation monitoring systems based on Internet of Things technology have emerged. Most of the existing systems use computer vision and image processing technology to monitor intersections in real time to detect vehicle violations. However, the existing technology still has some limitations in non-motor vehicle classification, personal information identification, and action trajectory analysis, such as the difficulty in accurately distinguishing between electric bicycles and bicycles, the difficulty in tracking and recording illegal vehicles, and the privacy issues in the information recording process. The existing technology is also insufficient in the continuous monitoring and behavior analysis of illegal vehicles, which limits the timely response and management of traffic violations.

[0003] Therefore, this patent aims to provide an intelligent traffic monitoring system based on image processing, face / license plate recognition and motion trajectory analysis to solve the problem of non-motor vehicle violations in a more comprehensive and efficient manner, thereby improving the accuracy and real-time performance of violation detection and reducing the adverse effects of violations on traffic safety and order. Summary of the invention

[0004] The purpose of the present invention is to provide an intelligent traffic monitoring method and system based on the Internet of Things to solve the problems raised in the above background technology.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: an intelligent traffic monitoring system based on the Internet of Things, the operation method of the system comprises the following steps:

[0006] Step 1: Distinguish non-motor vehicles by combining the angle between the seat and the human body and the driving speed;

[0007] Step 2: Detect non-motor vehicle violations and issue warnings for violations;

[0008] Step 3: Record the violation information of non-motor vehicles and generate corresponding feature information;

[0009] Step 4: Collect violation information and analyze movement trajectories of vehicles in the file.

[0010] According to the above technical solution, the step of distinguishing non-motor vehicles by combining the angle between the seat and the human body and the driving speed includes:

[0011] Calculate the angle between the seat and the back, as well as the speed of the non-motor vehicle;

[0012] Set the speed threshold and angle threshold, and compare them with the actual values ​​to identify the vehicle type.

[0013] According to the above technical solution, the step of calculating the angle between the seat and the back, and the speed of the non-motor vehicle includes:

[0014] The system is connected to a camera device. The system obtains image information during riding through the camera device. The camera device uses a high-resolution camera that can capture key parts of non-motor vehicles. When a non-motor vehicle passes under the camera device, the camera device takes a picture of the current shooting range and then uploads it to the system for image processing. The system first outlines the vehicle in the image by using the Canny edge detection algorithm to highlight the outline of the vehicle. Next, the angle between the seat and the back is calculated. The image processing technology is used to identify the seat and back positions of the vehicle. Then, the angle between the two is determined by vector angle calculation (θ): Among them, V 座位 and V 后背 They are the vectors of the seat and the back respectively. Then the speed of the non-motor vehicle is calculated. The video is shot by a camera, and then a series of continuous image frames are obtained through the video. Then an image sequence is established, in which each frame contains a non-motor vehicle in motion. At the same time, SIFT is used to detect feature points in each frame of the image sequence, and feature points are matched between adjacent frames. Then, the optical flow method or the motion estimation algorithm based on feature points is used to calculate the motion vector of the feature points between adjacent frames. Finally, the pixel speed of each feature point is calculated according to the time interval Δt between adjacent frames and the displacement Δx, Δy of the vehicle feature point in the image. For a certain feature point, the speed Finally, the speeds of all feature points are averaged, and the average speed of the vehicle is obtained by taking a weighted average of the speeds of all feature points.

[0015] According to the above technical solution, the step of setting the speed threshold and the angle threshold and comparing them with the actual values ​​to identify the vehicle type includes:

[0016] By setting the speed threshold and angle threshold, the electric bicycle and bicycle are identified. First, a speed threshold V is set according to the actual situation. 阈值 and an angle threshold θ 阈值 , used to determine the type of vehicle. The selection of these two thresholds can be determined through experiments. When the measured vehicle speed V is greater than V 阈值, it can be preliminarily determined to be an electric bicycle, because electric bicycles usually have a higher speed. Otherwise, proceed to the next step. If the speed cannot be determined, consider the angle θ. When the angle is less than θ 阈值 , it is determined to be an electric bicycle, because the riding posture of an electric bicycle is more upright and the angle is smaller. On the contrary, the angle is greater than θ 阈值 It's a bicycle.

[0017] According to the above technical solution, the steps of detecting the non-motor vehicle's illegal behavior and issuing a warning for the illegal behavior include:

[0018] Through image processing technology, the status of traffic lights at intersections is detected, and then the system determines whether non-motor vehicles have run red lights based on the status of traffic lights. At the same time, the position and movement trajectory of non-motor vehicles at intersections are analyzed. By tracking the position changes of non-motor vehicles in image sequences, their driving direction and whether they are traveling in the opposite direction can be inferred. For violations of electric bicycles, the system first obtains the corresponding facial information, and performs facial recognition by adopting a facial recognition system based on convolutional neural networks. Then, the identity information of the non-motor vehicle owner is obtained through facial information. Secondly, the license plate of the electric bicycle is recognized, the license plate information of the electric bicycle is extracted, and the personal information corresponding to the license plate information is obtained. After successfully identifying the violation of the electric bicycle and obtaining the personal information, the system automatically triggers the SMS warning system. Through SMS or phone calls, the system sends a text message containing the violation information to the owner of the electric bicycle. For violations of bicycles, the system first attempts to perform facial recognition. After success, the system records the violation information and issues a warning. When facial recognition fails, the system introduces a personal violation file module, which generates a unique identifier for each illegal bicycle, then associates the violation information with the identifier, and establishes a personal violation file in the file module.

[0019] According to the above technical solution, the step of recording the violation information of non-motor vehicles and generating corresponding characteristic information includes:

[0020] A detailed record of each bicycle violation is kept, including but not limited to the time, location, vehicle color, weather conditions, and road conditions. The time record will include the date and specific time, and the location record will include the specific intersection or location information. Secondly, various features are introduced. Through the meteorological data source, the weather conditions when the violation occurred are recorded, such as sunny, rainy or snowy days, to consider the impact of weather on riding behavior. At the same time, through image processing technology or sensors, the road conditions, such as slippery and potholes, are recorded. Finally, all violation information and related features of the bicycle are securely stored and protected by encryption algorithms to prevent unauthorized access.

[0021] According to the above technical solution, the steps of collecting traffic violation information and analyzing movement trajectories of vehicles in the archives include:

[0022] Start action trajectory analysis, using clustering algorithm and trajectory similarity calculation;

[0023] The prediction results are provided to the transportation department to increase supervision and reduce violations.

[0024] According to the above technical solution, the step of starting action trajectory analysis, using clustering algorithm and trajectory similarity calculation, includes:

[0025] The system counts the number of violations of the vehicle. Once the number of violations exceeds the preset threshold, the system will start the action trajectory analysis. In the action trajectory analysis stage, the system uses clustering algorithms and trajectory similarity calculations. First, the historical driving trajectory data of the vehicle, including location coordinates and timestamp information, is extracted from the bicycle information file, and data preprocessing is performed. Then, the system uses the density-based DBSCAN clustering algorithm to perform cluster analysis on the historical trajectory of the vehicle. By grouping similar trajectory points, the system identifies the places or intersections where the vehicle frequently appears in different time periods. At the same time, the system performs trajectory similarity calculations and uses the dynamic time warping method to perform detailed similarity evaluations on each category in the clustering results. By finding more similar trajectories in the same category, the distribution of vehicles in frequently visited places is further refined. Finally, the system combines the results of clustering analysis and trajectory similarity calculations to determine the intersections or areas where the vehicle frequently appears, and stores the information of these frequently visited places by setting thresholds or marking them using cluster centers.

[0026] According to the above technical solution, the system includes:

[0027] Non-motor vehicle classification and differentiation module, which is used to capture non-motor vehicle images in real time, accurately classify electric bicycles and bicycles through advanced image processing technology, and provide key movement information;

[0028] Traffic violation detection and warning module, which is used to accurately detect non-motor vehicle violations through traffic light status detection and face / license plate recognition, triggering the warning system to ensure traffic safety;

[0029] The violation information recording and feature adding module is used to record each bicycle violation in detail, including time, location, vehicle characteristics, add multiple features, and provide comprehensive and in-depth violation information.

[0030] According to the above technical solution, the non-motor vehicle classification and differentiation module includes:

[0031] Image information acquisition module, used to capture non-motor vehicle images through a high-resolution camera and provide key information to the system for processing;

[0032] Image processing and classification module, used to accurately distinguish between electric bicycles and bicycles using Canny edge detection and vector angle calculation;

[0033] The electric bicycle and bicycle recognition module is used to set speed and angle thresholds to accurately distinguish electric bicycles from bicycles, improving classification accuracy in real-time scenarios.

[0034] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: the present invention, first, distinguishes between electric bicycles and bicycles by combining the angle between the seat and the human body and the driving speed; then, in step two, the system detects non-motor vehicle violations, such as running a red light or driving in the wrong direction, obtains personal information through facial recognition, license plate recognition and other technologies, and issues warnings through text messages; step three records detailed violation information, including time, location, weather and other characteristics; finally, in step four, the system realizes the statistics of the number of violations and the marking of frequently visited intersections through statistics and trajectory analysis of vehicle information files, providing accurate prediction results for the transportation department to more effectively manage and supervise non-motor vehicle violations; this method has the characteristics of improving the accuracy and real-time performance of violation detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0036] Figure 1 A flow chart of an intelligent traffic monitoring method based on the Internet of Things provided in Embodiment 1 of the present invention;

[0037] Figure 2 A schematic diagram of the module composition of an intelligent traffic monitoring system based on the Internet of Things provided in Embodiment 2 of the present invention. DETAILED DESCRIPTION

[0038] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0039] Embodiment 1: Figure 1 Flow chart of the intelligent traffic monitoring method based on the Internet of Things provided in the first embodiment of the present invention. This embodiment can be applied to the scenario of non-motor vehicle violation monitoring. The method can be executed by the intelligent traffic monitoring system based on the Internet of Things provided in this embodiment. Figure 1As shown, the method specifically comprises the following steps:

[0040] Step 1: Distinguish non-motor vehicles by combining the angle between the seat and the human body and the driving speed;

[0041] In the embodiment of the present invention, the angle between the seat and the back and the driving speed are combined to accurately distinguish between electric bicycles and bicycles;

[0042] Exemplarily, the system is connected to a camera device, and the system obtains image information during riding through the camera device. The camera device uses a high-resolution camera that can capture key parts of non-motor vehicles. When the non-motor vehicle passes under the camera device, the camera device takes a picture of the current shooting range, and then uploads it to the system for image processing. The system first outlines the vehicle in the image by using the Canny edge detection algorithm to highlight the outline of the vehicle. Next, the angle between the seat and the back is calculated, and the seat and back positions of the vehicle are identified using image processing technology. Then, the angle (θ) between the two is determined by vector angle calculation: Among them, V 座位 and V 后背 They are the vectors of the seat and the back, respectively. Then the speed of the non-motor vehicle is calculated. The video is shot by a camera, and then a series of continuous image frames are obtained through the video. Then an image sequence is established, in which each frame contains a non-motor vehicle in motion. At the same time, SIFT is used to detect feature points in each frame of the image sequence, and feature points are matched between adjacent frames. Then, the optical flow method or the motion estimation algorithm based on feature points is used to calculate the motion vector of the feature points between adjacent frames. Finally, the pixel speed of each feature point is calculated according to the time interval Δt between adjacent frames and the displacement Δx, Δy of the vehicle feature point in the image. For a certain feature point, the speed v can be calculated by the following formula: Finally, the speeds of all feature points are averaged, and the average speed of the vehicle is obtained by weighted averaging the speeds of all feature points to improve the accuracy of the overall movement of the vehicle. Through the above steps, the system can accurately measure the actual speed of non-motor vehicles in the image sequence, providing key movement information for subsequent vehicle classification;

[0043] For example, the angle and speed can be calculated through the above steps, and then the electric bicycle and the bicycle can be identified by setting the speed threshold and the angle threshold. First, a speed threshold V is set according to the actual situation. 阈值 and an angle threshold θ 阈值 , used to determine the type of vehicle. The selection of these two thresholds can be determined through experiments. When the measured vehicle speed V is greater than V 阈值, it can be preliminarily determined to be an electric bicycle, because electric bicycles usually have a higher speed. Otherwise, proceed to the next step. If the speed cannot be determined, consider the angle θ. When the angle is less than θ 阈值 , it is determined to be an electric bicycle, because the riding posture of an electric bicycle is more upright and the angle is smaller. On the contrary, the angle is greater than θ 阈值 It is a bicycle. Through the above steps, combined with the speed and angle information, it is possible to accurately distinguish between bicycles and electric bicycles in real-time scenes.

[0044] Step 2: Detect non-motor vehicle violations and issue warnings for violations;

[0045] In the embodiment of the present invention, after distinguishing non-motor vehicles, violation identification is performed to check whether there is red light running or reverse driving, and then personal information is obtained for non-motor vehicles with illegal behaviors, and warnings are issued via SMS;

[0046] For example, the state of the traffic lights at the intersection is detected through image processing technology, and then the system determines whether the non-motor vehicle has run a red light based on the state of the traffic lights. At the same time, the position and movement trajectory of the non-motor vehicle at the intersection are analyzed. By tracking the position changes of the non-motor vehicle in the image sequence, its driving direction and whether it is going in the wrong direction can be inferred. For illegal behavior of electric bicycles, the system first obtains the corresponding facial information, and performs facial recognition by adopting a facial recognition system based on a convolutional neural network. Then, the identity information of the non-motor vehicle owner is obtained through the facial information. Secondly, the license plate of the electric bicycle is recognized, the license plate information of the electric bicycle is extracted, and the personal information corresponding to the license plate information is obtained. After successfully identifying the electric bicycle violation and obtaining personal information, the system automatically triggers the SMS warning system. The system sends a text message containing the violation information to the electric bicycle owner through SMS or phone calls. For bicycle violations, the system first attempts facial recognition. If successful, the system records the violation information and issues a warning. If facial recognition fails, the system introduces a personal violation file module, which generates a unique identifier for each illegal bicycle, and then associates the violation information with the identifier, and establishes a personal violation file in the file module. Through this step, the identification of electric bicycle and bicycle violation information can be accurately achieved, which is conducive to reducing violations when driving non-motor vehicles.

[0047] Step 3: Record the violation information of non-motor vehicles and generate corresponding feature information;

[0048] In the embodiment of the present invention, after the facial recognition fails, the bicycle is further registered for violation information, and more comprehensive violation information is provided by adding multiple features of the illegal vehicle;

[0049] Exemplarily, a detailed record is made of each bicycle violation, including but not limited to the time, location, vehicle color, weather conditions, road conditions, etc. The time record will include the date and specific time, and the location record will include the specific intersection or location information. Secondly, various features are introduced. Through the meteorological data source, the weather conditions when the violation occurred are recorded, such as sunny, rainy or snowy days, to consider the impact of weather on riding behavior. At the same time, through image processing technology or sensors, the road conditions, such as slippery, potholes, etc., are recorded. Finally, all violation information and related features of the bicycle are securely stored and protected by encryption algorithms to prevent unauthorized access. The feature information recording in this step can provide a more comprehensive and in-depth understanding of bicycle violation information, and continuous vehicle monitoring can be carried out through this information in the future.

[0050] Step 4: Collect violation information and analyze movement trajectories of vehicles in the file.

[0051] In the embodiment of the present invention, we monitor the bicycle information file, combine the violation statistics and action trajectory analysis, perform violation prediction and frequently visited intersection marking, so as to improve the management effect of traffic violation behavior;

[0052] Exemplarily, based on the violation information recorded in step three, the system counts the number of violations of the vehicle. Once the number of violations exceeds a preset threshold, the system will start the action trajectory analysis. In the action trajectory analysis stage, the system uses a clustering algorithm and trajectory similarity calculation. First, the historical driving trajectory data of the vehicle is extracted from the bicycle information file, including key information such as location coordinates and timestamps, and data preprocessing is performed. Then, the system uses a density-based DBSCAN clustering algorithm to perform cluster analysis on the historical trajectory of the vehicle. By grouping similar trajectory points, the locations or intersections where the vehicle frequently appears in different time periods are identified. At the same time, the system performs trajectory similarity calculation and uses a dynamic time warping method to perform a detailed similarity evaluation on each category in the clustering results. By finding more similar trajectories in the same category, the distribution of the vehicle in frequently visited locations is further refined. Finally, the system combines the results of cluster analysis and trajectory similarity calculation to determine the intersections or areas where the vehicle frequently appears, and stores the information of these frequently visited locations by setting thresholds or marking them using cluster centers.

[0053] For example, when the number of vehicle violations reaches a threshold, the system uses historical driving patterns to predict the next travel time and location. By using machine learning models or time series analysis methods, the system can accurately predict the vehicle's possible next travel path. Finally, the prediction results are provided to the transportation department. The transportation department goes to the predicted intersection based on the prediction results and increases supervision to effectively reduce vehicle violations. In this process, the system provides targeted prediction results by closely monitoring vehicle activities, making traffic management more efficient.

[0054] Embodiment 2: Embodiment 2 of the present invention provides an intelligent traffic monitoring system based on the Internet of Things. Figure 2 The schematic diagram of the module composition of the intelligent traffic monitoring system based on the Internet of Things provided in the second embodiment of the present invention is as follows: Figure 2 As shown, the system includes:

[0055] Non-motor vehicle classification and differentiation module, which is used to capture non-motor vehicle images in real time, accurately classify electric bicycles and bicycles through advanced image processing technology, and provide key movement information;

[0056] Traffic violation detection and warning module, which is used to accurately detect non-motor vehicle violations through traffic light status detection and face / license plate recognition, triggering the warning system to ensure traffic safety;

[0057] The violation information recording and feature adding module is used to record each bicycle violation in detail, including time, location, vehicle characteristics, add multiple features, and provide comprehensive and in-depth violation information;

[0058] In some embodiments of the present invention, the non-motor vehicle classification and differentiation module includes:

[0059] Image information acquisition module, used to capture non-motor vehicle images through a high-resolution camera and provide key information to the system for processing;

[0060] Image processing and classification module, used to accurately distinguish between electric bicycles and bicycles using Canny edge detection and vector angle calculation;

[0061] The electric bicycle and bicycle identification module is used to set speed and angle thresholds to accurately distinguish electric bicycles from bicycles, improving classification accuracy in real-time scenarios;

[0062] In some embodiments of the present invention, the violation detection and warning module includes:

[0063] Traffic light status detection module, used to detect the status of traffic lights through image processing and determine whether non-motor vehicles run red lights;

[0064] The electric bicycle violation processing module is used to obtain the owner's information of the electric bicycle using facial recognition and license plate recognition technology and trigger the SMS warning system;

[0065] The bicycle violation processing module is used to perform facial recognition and record violation information. If it fails, the personal violation file is introduced to record it, so as to improve the recognition accuracy of bicycle violations;

[0066] In some embodiments of the present invention, the violation information recording and feature adding module includes:

[0067] The violation information recording module is used to record each bicycle violation in detail, including time, location, vehicle characteristics, etc., and provide comprehensive violation information;

[0068] Feature information adding module, used to introduce meteorological data and road condition records, enrich violation information, and consider the impact of weather and road conditions on riding behavior;

[0069] The vehicle monitoring and statistical analysis module is used to statistically analyze the number of vehicle violations, initiate movement trajectory analysis, identify frequently visited locations, and provide the transportation department with a forecast of the next travel time and location to improve management effectiveness.

[0070] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0071] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An intelligent traffic monitoring method based on the Internet of Things, characterized in that: The method comprises the following steps: Step 1: Distinguish non-motor vehicles by combining the angle between the seat and the human body and the driving speed; Step 2: Detect non-motor vehicle violations and issue warnings for violations; Step 3: Record the violation information of non-motor vehicles and generate corresponding feature information; Step 4: Analyze the traffic violation information and movement trajectory of the vehicles in the file; The step of distinguishing non-motor vehicles by combining the angle between the seat and the human body and the driving speed includes: Calculate the angle between the seat and the back, as well as the speed of the non-motor vehicle; Set speed threshold and angle threshold, and compare them with actual values ​​to identify vehicle type; The step of calculating the angle between the seat and the back, and the speed of the non-motor vehicle comprises: The system is connected to a camera device. The system obtains image information during riding through the camera device. The camera device uses a high-resolution camera that captures the key parts of non-motor vehicles. When a non-motor vehicle passes under the camera device, the camera device takes a picture of the current shooting range and then uploads it to the system for image processing. The system first outlines the vehicle in the image by using the Canny edge detection algorithm to highlight the outline of the vehicle. Next, the angle between the seat and the back is calculated. The image processing technology is used to identify the seat and back positions of the vehicle. Then, the angle between the two is determined by vector angle calculation (θ): Among them, V 座位 and V 后背 They are the vectors of the seat and the back respectively. Then the speed of the non-motor vehicle is calculated. The video is shot by a camera, and then a series of continuous image frames are obtained through the video. Then an image sequence is established, in which each frame contains a non-motor vehicle in motion. At the same time, SIFT is used to detect feature points in each frame of the image sequence, and feature points are matched between adjacent frames. Then, the optical flow method or the motion estimation algorithm based on feature points is used to calculate the motion vector of the feature points between adjacent frames. Finally, the pixel speed of each feature point is calculated according to the time interval Δt between adjacent frames and the displacement Δx, Δy of the vehicle feature point in the image. For a certain feature point, the speed Finally, the speeds of all feature points are averaged, and the average speed of the vehicle is obtained by taking a weighted average of the speeds of all feature points; The step of setting a speed threshold and an angle threshold, and comparing them with actual values ​​to identify the vehicle type includes: By setting the speed threshold and angle threshold, the electric bicycle and bicycle are identified. First, a speed threshold V is set according to the actual situation. 阈值 and an angle threshold θ 阈值 , used to determine the type of vehicle. The selection of these two thresholds is determined by experiments. When the measured vehicle speed V is greater than V 阈值 , it is initially determined to be an electric bicycle. Otherwise, proceed to the next step. If the speed cannot be determined, consider the angle θ. When the angle is less than θ 阈值 , it is determined to be an electric bicycle. The riding posture of an electric bicycle is more upright and the angle is smaller. On the contrary, the angle is greater than θ 阈值 It's a bicycle.

2. The method for intelligent traffic monitoring based on the Internet of Things according to claim 1, characterized in that: The steps of detecting the non-motor vehicle's illegal behavior and issuing a warning for the illegal behavior include: Through image processing technology, the status of traffic lights at intersections is detected, and then the system determines whether non-motor vehicles have run red lights based on the status of traffic lights. At the same time, the position and movement trajectory of non-motor vehicles at intersections are analyzed. By tracking the position changes of non-motor vehicles in image sequences, their driving direction and whether they are traveling in the opposite direction are inferred. For violations of electric bicycles, the system first obtains the corresponding facial information, and performs facial recognition by adopting a facial recognition system based on convolutional neural networks. Then, the identity information of the non-motor vehicle owner is obtained through the facial information. Secondly, the license plate of the electric bicycle is recognized, the license plate information of the electric bicycle is extracted, and the personal information corresponding to the license plate information is obtained. After successfully identifying the violation of the electric bicycle and obtaining the personal information, the system automatically triggers the SMS warning system. Through SMS or phone calls, the system sends a text message containing the violation information to the owner of the electric bicycle. For violations of bicycles, the system first attempts to perform facial recognition. After success, the system records the violation information and issues a warning. When facial recognition fails, the system introduces a personal violation file module, which generates a unique identifier for each illegal bicycle, then associates the violation information with the identifier, and establishes a personal violation file in the file module.

3. The method for intelligent traffic monitoring based on the Internet of Things according to claim 2 is characterized in that: The step of recording the violation information of the non-motor vehicle and generating corresponding characteristic information includes: A detailed record of each bicycle violation is kept, including but not limited to the time, location, vehicle color, weather conditions, and road conditions. The time record will include the date and specific time, and the location record will include specific intersection or location information. Secondly, various features are introduced. Through meteorological data sources, the weather conditions when the violation occurred are recorded to consider the impact of weather on cycling behavior. At the same time, the road conditions are recorded through image processing technology or sensors. Finally, all violation information and related features of the bicycle are securely stored and protected by encryption algorithms to prevent unauthorized access.

4. The method for intelligent traffic monitoring based on the Internet of Things according to claim 3 is characterized in that: The steps of collecting traffic violation information and analyzing movement trajectories of vehicles in the archives include: Start action trajectory analysis, using clustering algorithm and trajectory similarity calculation; The prediction results are provided to the transportation department to increase supervision and reduce violations.

5. The method for intelligent traffic monitoring based on the Internet of Things according to claim 4 is characterized in that: The step of starting action trajectory analysis, using a clustering algorithm and trajectory similarity calculation, includes: The system counts the number of violations of the vehicle. Once the number of violations exceeds the preset threshold, the system will start the action trajectory analysis. In the action trajectory analysis stage, the system uses clustering algorithms and trajectory similarity calculations. First, the historical driving trajectory data of the vehicle, including location coordinates and timestamp information, is extracted from the bicycle information file, and data preprocessing is performed. Then, the system uses the density-based DBSCAN clustering algorithm to perform cluster analysis on the historical trajectory of the vehicle. By grouping similar trajectory points, the system identifies the places or intersections where the vehicle frequently appears in different time periods. At the same time, the system performs trajectory similarity calculations and uses the dynamic time warping method to perform detailed similarity evaluations on each category in the clustering results. By finding more similar trajectories in the same category, the distribution of vehicles in frequently visited places is further refined. Finally, the system combines the results of clustering analysis and trajectory similarity calculations to determine the intersections or areas where the vehicle frequently appears, and stores the information of these frequently visited places by setting thresholds or marking them using cluster centers.

Images