Traffic monitoring system based on machine vision

Abstract

The invention relates to the technical field of traffic management, and particularly discloses a traffic monitoring system based on machine vision, and the system comprises a partitioning module which is used for determining a target building in a preset monitoring region, and determining an initial region according to the position of the target building; the processing module is used for determining the vehicle density in the initial area and determining the congestion radius according to the vehicle density; determining a theoretical center according to the position of the vehicle in the crowding radius, and correcting the crowding radius according to the theoretical center to obtain a standard radius; drawing a curve of the standard radius changing along with time, calculating a target radius and determining a target area; and the prompting module is used for determining a fork road closest to the boundary of the target area on the to-be-determined route and sending early warning information when the vehicle arrives at the fork road. The driving route of the vehicle can be prompted according to the vehicle density, and noise pollution caused by too high vehicle density is avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of traffic management, and in particular to a traffic monitoring system based on machine vision. Background Art

[0002] Machine vision is a technology that enables computers or machines to have "visual" capabilities. It aims to simulate the functions of the human visual system through image processing, pattern recognition and data analysis to perceive, understand and analyze image or video content. Its core principle is to convert image data into processable digital signals, process them through software and algorithms, and identify the features and patterns therein.

[0003] Noise refers to the sound produced by a sound-generating body when it vibrates irregularly. Sound is generated by the vibration of an object and propagates in a certain medium (such as solid, liquid, gas) in the form of waves. Noise pollution is usually caused by humans. From a physiological point of view, any sound that interferes with people's rest, study and work, as well as the sound you want to hear, that is, unnecessary sound, is collectively called noise. When noise has a negative impact on people and the surrounding environment, it forms noise pollution.

[0004] Vehicles usually produce noise while driving. This noise mainly comes from factors such as the engine, exhaust system, friction between wheels and the road, and aerodynamics. In areas with heavy traffic, the noise generated by vehicles easily accumulates to form noise pollution, causing trouble to residents' lives and work. Summary of the invention

[0005] The purpose of the present invention is to provide a traffic monitoring system based on machine vision to solve the following technical problems:

[0006] Vehicles usually produce noise while driving. This noise mainly comes from factors such as the engine, exhaust system, friction between wheels and the road, and aerodynamics. In areas with heavy traffic, the noise generated by vehicles easily accumulates to form noise pollution, causing trouble to residents' lives and work.

[0007] The purpose of the present invention can be achieved through the following technical solutions:

[0008] A traffic monitoring system based on machine vision, comprising:

[0009] Partitioning module: set the monitoring area, determine the target building in the monitoring area, the target building includes but is not limited to schools, hospitals and residential areas, and draw a circle with the location of the target building as the center and a preset radius r to obtain the initial area;

[0010] Processing module: obtain the location information of the vehicles in the monitoring area, determine the vehicle density P in the initial area, set the vehicle density threshold Pys, and when the vehicle density P≥Pys, set the radius r=r+Δr and redetermine the initial area, Δr is the preset radius adjustment value, repeat the above steps until the vehicle density P<Pys after a certain radius adjustment, and mark the radius at this time as the congestion radius R;

[0011] Determine an initial area A corresponding to the congestion radius R, determine a theoretical center G according to the positions of vehicles in the initial area A, the theoretical center is the average distribution center of vehicles, and correct the congestion radius according to the distance between the position of the target building and the theoretical center G to obtain a standard radius;

[0012] The standard radius is periodically obtained and a curve f(t) of the standard radius changing with time is drawn, t∈[Tcur-n*ΔT, Tcur], Tcur represents the current time, ΔT is a preset time interval, n is a preset number, and the target radius Rgoa=f(Tcur+ΔT) is calculated. A circle is drawn with the target building location as the center and the target radius Rgoa to obtain the target area;

[0013] Prompt module: take the route with intersection between the boundary of the monitoring area and the target area as the pending route, determine the fork B on the pending route that is closest to the boundary of the target area, and send warning information to the vehicle when the vehicle arrives at the fork B.

[0014] As a further solution of the present invention: in the processing module, the process of correcting the crowding radius to obtain the standard radius specifically includes:

[0015] Determine the position coordinates Wi(Xi, Yi) of the i-th vehicle in the initial area A, where Xi and Yi are the horizontal and vertical coordinates of the i-th vehicle respectively, and determine the theoretical center G(Xthe, Ythe), where N represents the total number of vehicles in the initial area A;

[0016] Determine the distance D between the theoretical center G and the location of the target building, and calculate the standard radius RBZ=R-ε*D, where ε is a preset correction coefficient.

[0017] As a further solution of the present invention: in the processing module, the process of determining the crowding radius further includes the following steps:

[0018] The maximum radius rmax is set. During the process of adjusting the radius r, if the radius r=rmax and the corresponding vehicle density is still greater than or equal to the vehicle density threshold Pys, an early warning message is sent to prompt the management personnel, and the preset reduction steps are executed to determine the congestion radius.

[0019] As a further solution of the present invention: the reduction step specifically includes:

[0020] Let the radius r = r-Δr and redetermine the initial area until the radius r < rmin, where rmin represents the preset minimum radius, and determine the density value Hj. When the vehicle density P ≥ Pys corresponding to the jth radius adjustment, Hj = 1; when the vehicle density P < Pys corresponding to the jth radius adjustment, Hj = 0;

[0021] Sort the density values ​​in the order of the time axis, starting from the first position in the sorting, determine the sorting position c where the last density value of 1 is located in the sorting, determine the target position c', and use the radius corresponding to the target position c' as the crowding radius. The target position c' satisfies the constraint that the density values ​​in the sorting position interval [c', c] are all 1;

[0022] Among them, when there is no density value of 1 in the sorting, it is marked as a smooth state, and the congestion radius at this time is set to 0.

[0023] As a further solution of the present invention: in the processing module, when the vehicle density P in the initial area is less than Pys, the reduction step is executed to determine the congestion radius.

[0024] As a further solution of the present invention: when there is no sorting position that meets the constraint conditions, the radius corresponding to the sorting position c is used as the congestion radius.

[0025] As a further solution of the present invention: in the processing module, the process of determining the vehicle density P specifically includes:

[0026] P = a / (πr 2 ), a represents the total number of vehicles in the initial area.

[0027] As a further solution of the present invention: in the processing module, the position information of the vehicle is obtained based on the target detection algorithm.

[0028] Beneficial effects of the present invention: In the present invention, the target building in the monitoring area is first determined, and the initial area is determined according to the location of the target building; the target buildings (such as schools, hospitals, residential areas, etc.) are because these places are more sensitive to noise pollution, and noise will directly affect people's health, learning efficiency, rest quality, etc. The initial area is to define a potential noise interference range and provide a dynamically adjustable basis for subsequent analysis. The radius r of this initial range is not a fixed value, but a benchmark value. This preset initial area is mainly for the subsequent detection and dynamic adjustment of vehicle density to provide a basis. Based on the spatial correlation between vehicle distribution and noise sources, a fixed initial area can help simplify analysis and reduce monitoring workload; Pys (density threshold) is set as a scale to distinguish whether the noise impact exceeds the tolerance range of the area. When the vehicle density exceeds Pys, it indicates that the noise pollution may reach or exceed the acceptance limit of the building area. In this case, the radius r is expanded and monitoring continues, so that the area can contain more vehicles and better analyze the overall density. In this way, by dynamically adjusting the monitoring range, real-time monitoring is effectively ensured. Adjust the noise monitoring of the building; after calculating the theoretical center of the vehicle, use the distance D from the target building to the theoretical center as a correction factor to measure the noise impact range, because when the center of gravity of the vehicle distribution is close to the target building, the noise impact is greater; conversely, when the theoretical center is relatively far from the target building, the noise intensity will decay due to the distance. This correction helps to avoid invalid monitoring. When the vehicle distribution is far away from the target building but dense, this distribution has less impact on the noise of the building; then, according to the historical change trend of the standard radius, the target radius is obtained and the target area is determined, that is, the standard radius when the prediction time is Tcur+ΔT, and the time Tcur+ΔT is substituted into the curve f(t); this prediction method can help managers formulate practical prevention and control measures in advance, and avoid unreasonable prevention and control situations by dynamically adjusting the prevention and control measures; finally, determine the fork in the road on the pending route that is closest to the boundary of the target area, and send an early warning message when the vehicle reaches the fork in the road; this allows the vehicle to be prompted when it is close to the noise-sensitive area, so that the vehicle can adjust its driving strategy in advance, thereby effectively reducing the risk of noise propagation to the target area. The present invention can prompt the driving route of the vehicle according to the vehicle density, thereby avoiding noise pollution caused by excessively high vehicle density. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The present invention will be further described below in conjunction with the accompanying drawings.

[0030] Figure 1 The present invention is a flow chart of a traffic monitoring system based on machine vision. DETAILED DESCRIPTION

[0031] 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.

[0032] See also Figure 1 As shown, the present invention is a traffic monitoring system based on machine vision, comprising:

[0033] Partitioning module: set the monitoring area, determine the target building in the monitoring area, the target building includes but is not limited to schools, hospitals and residential areas, and draw a circle with the location of the target building as the center and a preset radius r to obtain the initial area;

[0034] Processing module: obtain the location information of the vehicles in the monitoring area, determine the vehicle density P in the initial area, set the vehicle density threshold Pys, and when the vehicle density P≥Pys, set the radius r=r+Δr and redetermine the initial area, Δr is the preset radius adjustment value, repeat the above steps until the vehicle density P<Pys after a certain radius adjustment, and mark the radius at this time as the congestion radius R;

[0035] Determine an initial area A corresponding to the congestion radius R, determine a theoretical center G according to the positions of vehicles in the initial area A, the theoretical center is the average distribution center of vehicles, and correct the congestion radius according to the distance between the position of the target building and the theoretical center G to obtain a standard radius;

[0036] The standard radius is periodically obtained and a curve f(t) of the standard radius changing with time is drawn, t∈[Tcur-n*ΔT, Tcur], Tcur represents the current time, ΔT is a preset time interval, n is a preset number, and the target radius Rgoa=f(Tcur+ΔT) is calculated. A circle is drawn with the target building location as the center and the target radius Rgoa to obtain the target area;

[0037] Prompt module: take the route with intersection between the boundary of the monitoring area and the target area as the pending route, determine the fork B on the pending route that is closest to the boundary of the target area, and send warning information to the vehicle when the vehicle arrives at the fork B.

[0038] It should be noted that the target building in the monitoring area is first determined, and the initial area is determined according to the location of the target building; the target buildings (such as schools, hospitals, residential areas, etc.) are because these places are more sensitive to noise pollution, and noise will directly affect people's health, learning efficiency, rest quality, etc. The initial area is to define a potential noise interference range and provide a dynamically adjusted basis for subsequent analysis. The radius r of this initial range is not a fixed value, but a benchmark value. This preset initial area is mainly to provide a basis for subsequent detection and dynamic adjustment of vehicle density. Based on the spatial correlation between vehicle distribution and noise sources, a fixed initial area can help simplify analysis and reduce monitoring workload; Pys (density threshold) is set as a scale to distinguish whether the noise impact exceeds the tolerance range of the area. When the vehicle density exceeds Pys, it indicates that the noise pollution may reach or exceed the acceptance limit of the building area. In this case, the radius r will be expanded and monitoring will continue so that the area can contain more vehicles and better analyze the overall density. In this way, by dynamically adjusting the monitoring range, real-time adjustment of the building is effectively ensured. Noise monitoring of buildings; after calculating the theoretical center of the vehicle, the distance D from the target building to the theoretical center is used as a correction factor to measure the noise impact range, because when the center of gravity of the vehicle distribution is close to the target building, the noise impact is greater; conversely, when the theoretical center is relatively far from the target building, the noise intensity will decay due to the distance. This correction helps to avoid invalid monitoring. When the vehicle distribution is far away from the target building but dense, this distribution has little impact on the noise of the building; then, according to the historical change trend of the standard radius, the target radius is obtained and the target area is determined, that is, the standard radius when the prediction time is Tcur+ΔT, and the time Tcur+ΔT is substituted into the curve f(t); this prediction method can help managers formulate practical prevention and control measures in advance, and avoid unreasonable prevention and control situations by dynamically adjusting the prevention and control measures; finally, determine the fork in the road on the pending route that is closest to the boundary of the target area, and send an early warning message when the vehicle reaches the fork in the road; this allows the vehicle to be prompted when it is close to the noise-sensitive area, so that the vehicle can adjust its driving strategy in advance, thereby effectively reducing the risk of noise propagation to the target area.

[0039] In another preferred implementation of the present invention, in the processing module, the process of correcting the crowding radius to obtain the standard radius specifically includes:

[0040] Determine the position coordinates Wi(Xi, Yi) of the i-th vehicle in the initial area A, where Xi and Yi are the horizontal and vertical coordinates of the i-th vehicle respectively, and determine the theoretical center G(Xthe, Ythe), where N represents the total number of vehicles in the initial area A;

[0041] Determine the distance D between the theoretical center G and the location of the target building, and calculate the standard radius RBZ=R-ε*D, where ε is a preset correction coefficient.

[0042] It is worth noting that the core purpose of this process is to ensure that the noise monitoring range more accurately reflects the impact of actual noise on the target building. The initial "crowding radius" is based on vehicle density, but does not fully take into account the specific distribution of vehicles and the actual distance of the noise source from the building. This correction allows the system to dynamically adjust the monitoring range to adapt to different vehicle distribution situations, thereby reducing unnecessary monitoring burdens; the standard radius RBZ obtained through correction is an optimization of the initial noise radius, making the monitoring results more accurate and effective, and reducing inefficient judgments on the impact of noise on target buildings.

[0043] In another preferred implementation of the present invention, in the processing module, the process of determining the crowding radius further includes the following steps:

[0044] The maximum radius rmax is set. During the process of adjusting the radius r, if the radius r=rmax and the corresponding vehicle density is still greater than or equal to the vehicle density threshold Pys, an early warning message is sent to prompt the management personnel, and the preset reduction steps are executed to determine the congestion radius.

[0045] It is understandable that setting the maximum radius rmax can avoid over-expansion of the monitoring area. When the radius reaches rmax, if the vehicle density is still higher than the threshold Pys, it indicates that the vehicles in the area are very dense and the noise has exceeded the management limit. At this time, the system sends an early warning message to notify the management personnel to take additional measures, such as diversion or restriction, to alleviate noise pollution.

[0046] In another preferred implementation of the present invention, the reduction step specifically includes:

[0047] Let the radius r = r-Δr and redetermine the initial area until the radius r < rmin, where rmin represents the preset minimum radius, and determine the density value Hj. When the vehicle density P ≥ Pys corresponding to the jth radius adjustment, Hj = 1; when the vehicle density P < Pys corresponding to the jth radius adjustment, Hj = 0;

[0048] Sort the density values ​​in the order of the time axis, starting from the first position in the sorting, determine the sorting position c where the last density value of 1 is located in the sorting, determine the target position c', and use the radius corresponding to the target position c' as the crowding radius. The target position c' satisfies the constraint that the density values ​​in the sorting position interval [c', c] are all 1;

[0049] Among them, when there is no density value of 1 in the sorting, it is marked as a smooth state, and the congestion radius at this time is set to 0.

[0050] It should be noted that the density values ​​between the sorting position intervals [c', c] are all 1. This is to ensure that the current monitoring radius setting can cover an area of ​​continuous high density rather than a single point of density (that is, a high density value is not suddenly detected when the radius is adjusted). By finding this stable area at the current time, the system can determine the most appropriate congestion radius; that is, when adjusting the monitoring radius in an area with dense vehicles, the effective monitoring range can be determined based on the high density values ​​of adjacent intervals (rather than a single point of high value detected by chance), which helps to avoid over-expansion or reduction of the monitoring area.

[0051] In another preferred implementation of the present invention, in the processing module, when the vehicle density P in the initial area is less than Pys, the reduction step is executed to determine the congestion radius.

[0052] In another preferred implementation of the present invention, when there is no sorting position that satisfies the constraint condition, the radius corresponding to the sorting position c is used as the crowding radius.

[0053] In another preferred implementation of the present invention, in the processing module, the process of determining the vehicle density P specifically includes:

[0054] P = a / (πr 2 ), a represents the total number of vehicles in the initial area.

[0055] In another preferred implementation of the present invention, in the processing module, the position information of the vehicle is obtained based on a target detection algorithm.

[0056] It is worth noting that the target detection algorithm can accurately identify and locate each vehicle in the monitoring area in real time. Obtaining the vehicle's location information in this way can intuitively understand the distribution of vehicles, which is convenient for monitoring density and spatial distribution characteristics of noise sources; target detection algorithms, especially modern deep learning algorithms such as YOLO, Faster R-CNN, etc., can identify vehicles in complex environments and even locate vehicles in congested or fast-moving situations.

[0057] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A traffic monitoring system based on machine vision, characterized in that: include: Partitioning module: set the monitoring area, determine the target building in the monitoring area, the target building includes but is not limited to schools, hospitals and residential areas, and draw a circle with the location of the target building as the center and a preset radius r to obtain the initial area; Processing module: obtain the location information of the vehicles in the monitoring area, determine the vehicle density P in the initial area, set the vehicle density threshold Pys, and when the vehicle density P≥Pys, set the radius r=r+Δr and redetermine the initial area, Δr is the preset radius adjustment value, repeat the above steps until the vehicle density P<Pys after a certain radius adjustment, and mark the radius at this time as the congestion radius R; Determine an initial area A corresponding to the congestion radius R, determine a theoretical center G according to the positions of vehicles in the initial area A, the theoretical center is the average distribution center of vehicles, and correct the congestion radius according to the distance between the position of the target building and the theoretical center G to obtain a standard radius; The standard radius is periodically obtained and a curve f(t) of the standard radius changing with time is drawn, t∈[Tcur-n*ΔT, Tcur], Tcur represents the current time, ΔT is a preset time interval, n is a preset number, and the target radius Rgoa=f(Tcur+ΔT) is calculated. A circle is drawn with the target building location as the center and the target radius Rgoa to obtain the target area; Prompt module: take the route with intersection between the boundary of the monitoring area and the target area as the pending route, determine the fork B on the pending route that is closest to the boundary of the target area, and send warning information to the vehicle when the vehicle arrives at the fork B.

2. A machine vision-based traffic monitoring system according to claim 1, characterized in that: In the processing module, the process of correcting the crowding radius to obtain the standard radius specifically includes: Determine the position coordinates Wi(Xi, Yi) of the i-th vehicle in the initial area A, where Xi and Yi are the horizontal and vertical coordinates of the i-th vehicle respectively, and determine the theoretical center G(Xthe, Ythe), where N represents the total number of vehicles in the initial area A; Determine the distance D between the theoretical center G and the location of the target building, and calculate the standard radius RBZ=R-ε*D, where ε is a preset correction coefficient.

3. The machine vision-based traffic monitoring system according to claim 1, characterized in that: In the processing module, the process of determining the crowding radius further includes the following steps: The maximum radius rmax is set. During the process of adjusting the radius r, if the radius r=rmax and the corresponding vehicle density is still greater than or equal to the vehicle density threshold Pys, an early warning message is sent to prompt the management personnel, and the preset reduction steps are executed to determine the congestion radius.

4. A machine vision-based traffic monitoring system according to claim 3, characterized in that: The reduction steps specifically include: Let the radius r = r-Δr and redetermine the initial area until the radius r < rmin, where rmin represents the preset minimum radius, and determine the density value Hj. When the vehicle density P ≥ Pys corresponding to the jth radius adjustment, Hj = 1; when the vehicle density P < Pys corresponding to the jth radius adjustment, Hj = 0; Sort the density values ​​in the order of the time axis, starting from the first position in the sorting, determine the sorting position c where the last density value of 1 is located in the sorting, determine the target position c', and use the radius corresponding to the target position c' as the crowding radius. The target position c' satisfies the constraint that the density values ​​in the sorting position interval [c', c] are all 1; Among them, when there is no density value of 1 in the sorting, it is marked as a smooth state, and the congestion radius at this time is set to 0.

5. A machine vision-based traffic monitoring system according to claim 4, characterized in that: In the processing module, when the vehicle density P in the initial area is less than Pys, the reduction step is executed to determine the congestion radius.

6. A machine vision-based traffic monitoring system according to claim 4, characterized in that: When there is no sorting position that satisfies the constraint conditions, the radius corresponding to the sorting position c is used as the congestion radius.

7. The machine vision-based traffic monitoring system according to claim 1, characterized in that: In the processing module, the process of determining the vehicle density P specifically includes: P = a / (πr 2 ), a represents the total number of vehicles in the initial area.

8. The machine vision-based traffic monitoring system according to claim 1, characterized in that: In the processing module, the position information of the vehicle is obtained based on the target detection algorithm.

Images