Intelligent traffic congestion state monitoring method and system

Abstract

The invention relates to the technical field of traffic state monitoring, in particular to an intelligent traffic congestion state monitoring method and system, and the method comprises the steps: collecting a traffic road image of a road surface in real time; equally dividing the traffic road image into sub-images, and identifying vehicle areas in the sub-images; determining a local distribution coefficient of each vehicle area; obtaining a vehicle congestion area in each sub-image; analyzing the change trend of all the local distribution coefficients in each sub-image and the proportion of a vehicle congestion area, and determining the road complexity of each sub-image; determining a scale parameter in an image enhancement algorithm by using the road complexity, and performing image enhancement on the traffic road image; and based on the enhanced traffic road image, in combination with a neural network, obtaining a vehicle occupancy rate in the traffic road image, and determining a traffic congestion state at each moment. Therefore, the accuracy of traffic jam state monitoring is improved.

Description

Technical Field

[0001] The present application relates to the technical field of traffic status monitoring, and specifically to a smart traffic congestion status monitoring method and system. Background Art

[0002] With the improvement of living standards and the continuous growth of transportation demand, traffic congestion has become increasingly prominent, seriously affecting people's travel efficiency and the overall operation efficiency of the city, especially during special periods such as holidays. As a new type of traffic management model, smart transportation can better play the transportation efficiency of the transportation system and is of great significance to improving traffic operation efficiency and traffic safety.

[0003] In recent years, computer vision technology has developed rapidly, and image-based traffic monitoring methods have increasingly highlighted their unique advantages. With the help of high-definition cameras installed along the road to obtain real-time traffic images, traffic congestion can be automatically monitored. However, in the actual monitoring process, given the different levels and states of vehicle aggregation on different road sections, when using traditional image processing technology for image enhancement, due to differences in vehicle density and changes in driving modes, if the same scale parameters are used, it is difficult to accurately highlight the detailed features in the image, which in turn affects the accuracy of traffic congestion status monitoring. Summary of the invention

[0004] In order to solve the above technical problems, the purpose of this application is to provide a smart traffic congestion status monitoring method and system, and the technical solutions adopted are as follows:

[0005] In a first aspect, an embodiment of the present application provides a method for monitoring a smart traffic congestion state, the method comprising the following steps:

[0006] Collect traffic road images on the road in real time;

[0007] The process of monitoring traffic congestion status based on traffic road images is as follows:

[0008] The traffic road image is divided into sub-images, and each vehicle area in each sub-image is identified; the distance relationship and driving direction difference between each vehicle area and its adjacent vehicle areas are analyzed to obtain the local distribution coefficient of each vehicle area;

[0009] Based on the local distribution coefficients of all vehicle areas in each sub-image, determine the vehicle congestion area in each sub-image; analyze the change trend of all the local distribution coefficients in each sub-image and the proportion of the vehicle congestion area to determine the road complexity of each sub-image;

[0010] The road complexity is used to determine the scale parameter in the image enhancement algorithm, and the traffic road image is enhanced; based on the enhanced traffic road image, the traffic congestion state at each moment is determined.

[0011] In one embodiment, the determination of each vehicle area includes:

[0012] Edge detection and connected domain analysis are performed on each sub-image to obtain each connected domain, and each connected domain is used as each vehicle area.

[0013] In one embodiment, the determination of the local distribution coefficient includes:

[0014] For each sub-image, calculating the metric distance between the center coordinates of each vehicle region and its adjacent vehicle regions; calculating the sum of the areas of each vehicle region and its adjacent vehicle regions; calculating the ratio of the sum to the metric distance, and obtaining a fusion result of the ratios of each vehicle region and all its adjacent vehicle regions;

[0015] Determining a direction deviation of each vehicle region based on a deviation of the driving direction of each vehicle region from all adjacent vehicle regions;

[0016] The local distribution coefficient of each vehicle area is obtained by combining the fusion result and the direction deviation.

[0017] In one embodiment, the direction deviation is a normalized result of the average of the driving direction angles between each vehicle region and all its adjacent vehicle regions.

[0018] In one embodiment, the local distribution coefficient is the product of the fusion result and the directional deviation.

[0019] In one embodiment, the determining of the vehicle congestion area includes:

[0020] The local distribution coefficients of all vehicle areas in each sub-image are divided into categories using a clustering algorithm, and the mean of all the local distribution coefficients in each category is calculated. The vehicle areas in the category with the maximum mean value are recorded as vehicle congestion areas.

[0021] In one embodiment, the determination of the road complexity includes:

[0022] Calculate the average value of all the local distribution coefficients in each sub-image, obtain the proportion of the vehicle congestion area in each sub-image in all vehicle areas, and obtain a fitting straight line after all the local distribution coefficients in each sub-image are arranged in ascending order;

[0023] The road complexity is positively correlated with the average value and the proportion, and negatively correlated with the slope of the fitting straight line.

[0024] In one embodiment, the scale parameter is a negative correlation mapping result of the road complexity.

[0025] In one embodiment, determining the traffic congestion status at each time includes:

[0026] Based on the enhanced traffic road image and combined with the neural network, the vehicle occupancy rate in the traffic road image is obtained, specifically:

[0027] For the enhanced traffic road image at each moment, the ratio of the number of pixels in all vehicle areas to the number of all pixels in the traffic road image is calculated as the vehicle occupancy rate on the road at each moment;

[0028] If the vehicle occupancy rate is greater than or equal to a preset threshold, the traffic state at the corresponding moment is determined to be congested; otherwise, it is determined to be not congested.

[0029] In the second aspect, an embodiment of the present application also provides an intelligent traffic congestion status monitoring system, including a memory, a processor, and a computer program stored in the memory and running on the processor, and when the processor executes the computer program, the steps of any one of the above methods are implemented.

[0030] This application has at least the following beneficial effects:

[0031] The present application acquires traffic road images of the road surface in real time; divides the traffic road images into sub-images, identifies each vehicle area in each sub-image; analyzes the distance relationship and driving direction difference between each vehicle area and its adjacent vehicle areas to obtain the local distribution coefficient of each vehicle area; the local distribution coefficient reflects the density of vehicles and the chaotic degree of vehicle driving in the local range of each vehicle area, and improves the reliability of determining the congestion degree of the vehicle area; further, based on the local distribution coefficients of all vehicle areas in each sub-image, determines the vehicle congestion area in each sub-image; analyzes the change trend of all the local distribution coefficients in each sub-image, as well as the proportion of vehicle congestion areas, and determines each The road complexity of the sub-images; the road complexity reflects the complexity of the vehicle area distribution in each sub-image, reflects the possibility of vehicle congestion in each sub-image, and improves the accuracy of vehicle congestion status judgment; the road complexity is used to determine the scale parameters in the image enhancement algorithm, and the traffic road image is enhanced; by flexibly adjusting the scale parameters in the image enhancement algorithm, the image enhancement effect is improved, the detail features in the traffic road image are accurately highlighted, and the judgment of the traffic congestion status is facilitated; based on the enhanced traffic road image, combined with the neural network, the vehicle occupancy rate in the traffic road image is obtained to determine the traffic congestion status at the moment; finally, the accuracy of traffic congestion status monitoring is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present application or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0033] Figure 1 A flowchart of a method for monitoring a smart traffic congestion state provided in accordance with an embodiment of the present application;

[0034] Figure 2 Determine the flow chart for the scale parameters. DETAILED DESCRIPTION

[0035] In order to further explain the technical means and effects adopted by the present application to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation method, structure, features and effects of a smart traffic congestion status monitoring method and system proposed in accordance with the present application, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0036] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0037] The following is a detailed description of a specific solution of a smart traffic congestion status monitoring method and system provided by the present application in conjunction with the accompanying drawings.

[0038] See also Figure 1 , which shows a flowchart of a method for monitoring a smart traffic congestion state provided by an embodiment of the present application, the method comprising the following steps:

[0039] S1, real-time collection of traffic road images on the road surface.

[0040] This embodiment monitors the congestion status of traffic roads in the city. First, high-definition cameras are evenly arranged along the road to capture traffic images of the traffic roads, wherein the capture time interval of the high-definition camera is 1 minute, and the traffic road image obtained is an RGB color image. In order to facilitate the monitoring of the congestion status of the traffic road and reduce the subsequent calculation cost, the obtained RGB color image of the traffic road is converted into a grayscale image by weighted averaging. The weighted averaging method is a well-known technology, and the specific process will not be repeated. The implementer can set the capture time interval according to the actual situation, and this embodiment does not limit it here.

[0041] S2, dividing the traffic road image into sub-images, identifying each vehicle area in each sub-image; analyzing the distance relationship and driving direction difference between each vehicle area and its adjacent vehicle areas, and obtaining the local distribution coefficient of each vehicle area.

[0042] In order to further analyze the distribution status of motor vehicles in a local area of ​​a traffic road image, this embodiment uses image segmentation technology to segment the collected traffic road image into N sub-images of equal size. In this embodiment, the number of N is set to 64, which can be set by the implementer according to the actual situation. This embodiment does not limit this. The image segmentation technology is well known in the art and will not be described in detail in this embodiment.

[0043] First, each sub-image of the traffic road image is taken as input, and the edge detection and connected domain analysis algorithms are used to obtain each connected domain in each sub-image, each connected domain is recorded as each vehicle area, and the center coordinates of the minimum circumscribed rectangle of each connected domain represent the location of each connected domain, wherein edge detection and connected domain analysis are both well-known technologies and will not be described in detail in this embodiment.

[0044] Secondly, due to the different sizes of different types of vehicles and the incomplete vehicles in the sub-images caused by image segmentation, this embodiment counts the number of pixels in each connected domain as the area of ​​each connected domain. The purpose of analyzing image features based on the distribution of connected domains in the image is to reflect the congestion state characteristics of the local area of ​​the road through the analysis of the connected domains in the local area, combined with the differences in the number of vehicles and vehicle types under traffic congestion.

[0045] Specifically, when a traffic road is in a congested state, vehicles have difficulty moving forward due to traffic congestion, so the vehicles in the traffic road image are densely distributed, the distance between vehicles is small, and vehicles move slowly or even stagnate. In addition, traffic congestion may also cause some motor vehicles to occupy non-motorized lanes, further increasing the motor vehicle occupancy rate of the road, that is, increasing the local distribution density between connected domains in the traffic road image.

[0046] Therefore, taking the nth sub-image in the traffic road image as an example, based on the K nearest neighbor search algorithm, the center position of each connected domain in the nth sub-image is used as input to obtain the nearest area of ​​each connected domain. In this embodiment, the number of nearest neighbor connected domains is set to K=8, and the implementer can set it according to the actual situation. And calculate the distance d between the center position of each connected domain and the center position of the nearest neighbor connected domain. When the traffic is congested, the distance between the vehicles is small, which makes the value of d smaller. However, due to the different sizes of vehicles on the traffic road, the sizes of the connected domains in the sub-images are inconsistent. When the vehicle is large, the area of ​​the connected domain corresponding to the motor vehicle will become larger, and then the distance d between the connected domains will become larger. Therefore, considering the local density of the local area in the traffic road image during actual driving and the regional change differences caused by the vehicle type, the congestion density of the local area is analyzed. Among them, the K nearest neighbor search algorithm is an existing well-known technology, and the specific process will not be repeated.

[0047] In addition, traffic congestion will make the vehicle's running trajectory unstable, and some motor vehicles will cut in and cut in. However, when the traffic road is in a normal traffic state, there are fewer vehicles on the road and the distance between vehicles is larger, which makes the vehicle occupancy rate on the road low, and motor vehicles move forward smoothly in the current lane, and rarely cut in and change lanes.

[0048] Based on the above analysis, this embodiment first analyzes the distribution state of local areas in the traffic road image. Considering that vehicles may cut in when there is traffic congestion, the cutting in will cause the motor vehicle to change its direction of travel, making the difference in the direction of travel between the cutting in and its nearest neighboring motor vehicle larger, thereby making the local area of ​​the traffic road image more complicated, that is, the cutting in and changing lane behavior will increase the difference in the main direction between the connected domain in the traffic road image and its nearest connected domain.

[0049] Therefore, taking the i-th connected domain in the n-th sub-image in the traffic road image as an example, this embodiment uses the optical flow method to detect each connected domain, that is, the driving direction of each vehicle, wherein the optical flow method for detecting the driving direction of the vehicle is an existing well-known technology, and this embodiment will not be described in detail here. The local distribution coefficient of each vehicle area is calculated, specifically:

[0050] For each sub-image, calculating the metric distance between the center coordinates of each vehicle region and its adjacent vehicle regions; calculating the sum of the areas of each vehicle region and its adjacent vehicle regions; calculating the ratio of the sum to the metric distance, and obtaining a fusion result of the ratios of each vehicle region and all its adjacent vehicle regions;

[0051] Based on the deviation of the driving direction of each vehicle area and all its adjacent vehicle areas, determining the direction deviation of each vehicle area; the direction deviation is the normalized result of the mean of the driving direction angles of each vehicle area and all its adjacent vehicle areas;

[0052] The local distribution coefficient of each vehicle area is obtained by combining the fusion result and the direction deviation.

[0053] It should be noted that fusion means combining multiple variables, which can be calculated by addition, multiplication, or a combination of addition and multiplication. This embodiment uses addition as the calculation method for fusion, and the Euclidean distance is used to measure the distance. The implementer can choose other existing feasible distance calculation methods according to actual conditions, and this embodiment does not limit this.

[0054] Specifically, the specific calculation method of the local distribution coefficient of each vehicle area is:

[0055] Where, MTC n,i is the local distribution coefficient of the ith vehicle area in the nth sub-image of the traffic road image, A n,i is the area of ​​the i-th vehicle region in the n-th sub-image in the traffic road image, A n,i,j is the area of ​​the jth adjacent vehicle region of the ith vehicle region in the nth sub-image of the traffic road image, d n,i,j is the metric distance between the ith vehicle region and its jth adjacent vehicle region in the nth sub-image in the traffic road image, K is the number of adjacent vehicle regions of the ith vehicle region in the nth sub-image in the traffic road image, and in this embodiment, K=8, is the mean of the driving direction angles between the ith vehicle region and all its adjacent vehicle regions in the nth sub-image in the traffic road image, and Norm() is the normalization function. is the driving direction deviation, is the fusion result.

[0056] It should be understood that when the local area of ​​the nth sub-image in the traffic road image is relatively complex, the distance d between the i-th connected domain and its j-th adjacent connected domain in the sub-image is n,i,j However, due to the different sizes of different connected domains, a larger connected domain will make the distance d between the centers of adjacent connected domains smaller. n,i,j Therefore, this embodiment uses the area between adjacent connected domains as a weight to correct the distance d caused by the large area of ​​the connected domain when the distance between adjacent connected domains is close. n,i,j The phenomenon of increasing, that is, the distance d between the i-th connected domain and its j-th adjacent connected domain n,i,j The smaller the connected domain area A n,i +An,i,j The larger the value is, the greater the local distribution density of the i-th connected domain is. At the same time, due to traffic congestion, some vehicles will cut in and cut in, which will change the forward direction between the i-th vehicle area and its adjacent vehicle area, that is, the calculated driving direction deviation of the i-th vehicle area becomes larger, which in turn makes the local distribution coefficient MTC of the i-th vehicle area of ​​the n-th sub-image of the traffic road image finally calculated n,i The bigger.

[0057] S3, based on the local distribution coefficients of all vehicle areas in each sub-image, determine the vehicle congestion area in each sub-image; analyze the change trend of all the local distribution coefficients in each sub-image, and the proportion of the vehicle congestion area, to determine the road complexity of each sub-image.

[0058] Furthermore, since the density of vehicles at different locations on the traffic road is different, that is, the local distribution coefficients corresponding to different connected domains in the nth sub-image are quite different, the local distribution coefficients of all vehicle areas in the nth sub-image of the traffic road image are arranged from small to large, and the arranged data is recorded as the data sequence α of the nth sub-image n . Take the data sequence α n As input, the least squares linear fit is used to obtain the slope b of the fitted line. n , and then calculate the data sequence α n The mean Slope of the fitted line b n The smaller the value is, the mean of the local distribution coefficients of all vehicle regions in the nth sub-image The larger it is, the more complex the distribution between vehicle areas in the nth sub-image is, and the greater the possibility of traffic congestion.

[0059] Further, the data sequence α n As input, K-means clustering is used for clustering, and the number of categories K is set to 2 to obtain the data sequence α n After clustering, the mean of all local distribution coefficients in each cluster is calculated, and the cluster with the largest mean is recorded as a crowded cluster, and the vehicle area corresponding to all local distribution coefficients in the crowded cluster is recorded as a vehicle congestion area, and the cluster with the smallest mean is recorded as a sparse cluster. Therefore, the more the number of crowded clusters, the denser the local distribution of the vehicle area in the nth sub-image. Therefore, this embodiment further calculates the number of data in the crowded cluster as a percentage of the data sequence α n The proportion of the total data volume n Among them, the least square method and the K-means clustering algorithm are both existing well-known technologies, and are not described in detail in this embodiment.

[0060] Based on the above analysis, the road complexity of each sub-image is calculated, and the road complexity and the mean The ratios are all positively correlated and negatively correlated with the slope of the fitting line. The specific calculation method is:

[0061] Where TCBR n is the road complexity of the nth sub-image in the traffic road image, is the mean value of the local distribution coefficients of all vehicle regions in the nth sub-image of the traffic road image, r n is the number of data in the crowded cluster of the nth sub-image in the traffic road image, accounting for the data sequence α n The proportion of the total data volume, b n is the slope of the fitting line of the data sequence of the nth sub-image in the traffic road image, β is a preset value greater than 0 to avoid the denominator being 0. In this embodiment, β=0.01, and the implementer can set it according to the actual situation, and this embodiment does not limit it here.

[0062] It should be understood that when traffic congestion occurs in the nth sub-image of the traffic road image, the local distribution density between the vehicle areas in the nth sub-image increases, thereby making the calculated average value of the local distribution coefficients corresponding to all vehicle areas in the nth sub-image increases, and the difference between the local distribution coefficients corresponding to each vehicle area in the nth sub-image is small, so that the calculated data sequence α n The slope of the fitted line is b n The smaller the value, the denser the local distribution of vehicle areas in the nth sub-image, and the larger the proportion of vehicle areas in the crowded clusters of the nth sub-image; thus, the calculated road complexity TCBR in the nth sub-image n The bigger.

[0063] S4, using the road complexity to determine the scale parameter in the image enhancement algorithm, and performing image enhancement on the traffic road image; based on the enhanced traffic road image, combined with a neural network, obtaining the vehicle occupancy rate in the traffic road image, and determining the traffic congestion status at each moment.

[0064] Accordingly, the road complexity of each sub-image in the traffic road image can be calculated. The road complexity fully considers the local distribution state between vehicle areas in the sub-image and the difference between the distribution of vehicle areas at different positions in the road. Therefore, the complexity of different areas in the traffic road image can be accurately reflected according to the road complexity.

[0065] Furthermore, this embodiment performs filtering processing on the traffic road image based on the road complexity and combined with the single-scale SSR algorithm. Since the greater the road complexity of the sub-image in the traffic road image, the more complex the local area of ​​the sub-image. Therefore, when the road complexity TCBR corresponding to the nth sub-image n When the size is larger, a smaller scale parameter σ should be set to highlight the local features of the image and retain more details. On the contrary, when the road complexity TCBR corresponding to the nth sub-image is n The smaller it is, the less complex the road is, and the corresponding traffic state is relatively smooth. At this time, the scale parameter can be adjusted to reduce it compared to when the road complexity is high, thereby highlighting the global characteristics of the image.

[0066] In addition, in the single-scale SSR algorithm of this embodiment, the value range of the scale parameter σ is 80-100, and the implementer can set the value range of the scale parameter according to the actual situation. Therefore, this embodiment performs a negative correlation mapping on the calculated road complexity of each sub-image, specifically obtaining the inverse of the road complexity of each sub-image, and mapping the inverse to between 80-100 through a linear transformation method to obtain a scale parameter, and using the obtained scale parameter in combination with the single-scale SSR algorithm, each sub-image is enhanced, and finally, an image-enhanced traffic road image is obtained. Among them, the single-scale SSR algorithm and the linear transformation method are both existing well-known technologies, and this embodiment will not be described in detail here. The scale parameter determination flow chart is shown below. Figure 2 shown.

[0067] Furthermore, this embodiment monitors the congestion phenomenon of traffic roads, and uses a CNN neural network to extract each vehicle in the traffic road image after image enhancement, and obtains each vehicle area in the traffic road image after image enhancement. The loss function of the CNN neural network during training uses cross entropy, and the CNN neural network and the training process of the CNN neural network are both existing well-known technologies. This embodiment will not be described in detail here, and the implementer can choose other existing feasible neural networks.

[0068] For the enhanced traffic road image at each moment, the ratio of the number of pixels in all vehicle areas to the number of all pixels in the traffic road image is calculated as the vehicle occupancy rate on the road at each moment. If the vehicle occupancy rate is greater than or equal to the preset threshold T, the road traffic state at that moment is determined to be congested; otherwise, the road traffic state at that moment is determined to be non-congested.

[0069] In this embodiment, T=0.75, which can be set by the implementer according to the actual situation, and is not limited in this embodiment.

[0070] Based on the same inventive concept as the above method, an embodiment of the present application also provides an intelligent traffic congestion status monitoring system, including a memory, a processor, and a computer program stored in the memory and running on the processor, and when the processor executes the computer program, the steps of any one of the above-mentioned intelligent traffic congestion status monitoring methods are implemented.

[0071] It should be noted that the above sequence of the embodiments of the present application is for description only and does not represent the advantages and disadvantages of the embodiments. The above is a description of a specific embodiment of this specification. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0072] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

[0073] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present application should be included in the protection scope of the present application.

Claims

1. A smart traffic congestion status monitoring method, characterized in that: The method comprises the following steps: Collect traffic road images on the road in real time; The process of monitoring traffic congestion status based on traffic road images is as follows: The traffic road image is divided into sub-images, and each vehicle area in each sub-image is identified; the distance relationship and driving direction difference between each vehicle area and its adjacent vehicle areas are analyzed to obtain the local distribution coefficient of each vehicle area; Based on the local distribution coefficients of all vehicle areas in each sub-image, determine the vehicle congestion area in each sub-image; analyze the change trend of all the local distribution coefficients in each sub-image and the proportion of the vehicle congestion area to determine the road complexity of each sub-image; The road complexity is used to determine the scale parameter in the image enhancement algorithm, and the traffic road image is enhanced; based on the enhanced traffic road image, the traffic congestion state at each moment is determined.

2. A smart traffic congestion status monitoring method as claimed in claim 1, characterized in that: The determination of each vehicle area includes: Edge detection and connected domain analysis are performed on each sub-image to obtain each connected domain, and each connected domain is used as each vehicle area.

3. A smart traffic congestion status monitoring method as claimed in claim 1, characterized in that: The determination of the local distribution coefficient includes: For each sub-image, calculating the metric distance between the center coordinates of each vehicle region and its adjacent vehicle regions; calculating the sum of the areas of each vehicle region and its adjacent vehicle regions; calculating the ratio of the sum to the metric distance, and obtaining a fusion result of the ratios of each vehicle region and all its adjacent vehicle regions; Determining a direction deviation of each vehicle region based on a deviation of the driving direction of each vehicle region from all adjacent vehicle regions; The local distribution coefficient of each vehicle area is obtained by combining the fusion result and the direction deviation.

4. A smart traffic congestion status monitoring method as claimed in claim 3, characterized in that: The direction deviation is a normalized result of the average of the driving direction angles between each vehicle area and all its adjacent vehicle areas.

5. A smart traffic congestion status monitoring method as claimed in claim 3, characterized in that: The local distribution coefficient is the product of the fusion result and the directional deviation.

6. A smart traffic congestion status monitoring method as claimed in claim 1, characterized in that: The determination of the vehicle congestion area includes: The local distribution coefficients of all vehicle areas in each sub-image are divided into categories using a clustering algorithm, and the mean of all the local distribution coefficients in each category is calculated. The vehicle areas in the category with the maximum mean value are recorded as vehicle congestion areas.

7. A smart traffic congestion status monitoring method as claimed in claim 1, characterized in that: The determination of the road complexity includes: Calculate the average value of all the local distribution coefficients in each sub-image, obtain the proportion of the vehicle congestion area in each sub-image in all vehicle areas, and obtain a fitting straight line after all the local distribution coefficients in each sub-image are arranged in ascending order; The road complexity is positively correlated with the average value and the proportion, and negatively correlated with the slope of the fitting straight line.

8. The intelligent traffic congestion status monitoring method according to claim 1, characterized in that: The scale parameter is a negative correlation mapping result of the road complexity.

9. The intelligent traffic congestion status monitoring method according to claim 1, characterized in that: Determining the traffic congestion status at each time comprises: Based on the enhanced traffic road image and combined with the neural network, the vehicle occupancy rate in the traffic road image is obtained, specifically: For the enhanced traffic road image at each moment, the ratio of the number of pixels in all vehicle areas to the number of all pixels in the traffic road image is calculated as the vehicle occupancy rate on the road at each moment; If the vehicle occupancy rate is greater than or equal to a preset threshold, the traffic state at the corresponding moment is determined to be congested; otherwise, it is determined to be not congested.

10. A smart traffic congestion status monitoring system, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 9 are implemented.

Images