An intelligent urban planning and management system and method based on big data

Abstract

The present invention discloses an intelligent urban planning and management system and method based on big data, which relates to the technical field of big data and includes: obtaining the planar layout diagram of a target intersection and historical surveillance videos to obtain the turning proportion vectors corresponding to each lane; extracting target video segments therefrom according to the surveillance videos, and then calculating the current congestion degree of each lane group; obtaining the effective signals that allow vehicles on each lane to pass through the target intersection, and obtaining the timing coefficients of the effective signals corresponding to each lane; obtaining the historical timing data of the effective signals corresponding to each lane, and adjusting the timing data of the effective signals corresponding to each lane according to the timing coefficients. According to the actual situation of the intersection, the present invention rationally and specifically optimizes the timing of each signal on the traffic lights, provides a more convenient traffic environment for residents driving cars, improves the passing efficiency of the intersection, and alleviates traffic congestion.

Description

Technical Field

[0001] The present invention relates to the technical field of big data, and specifically to an intelligent urban planning management system and method based on big data. Background Art

[0002] Urban planning management plays an important role in ensuring the orderly progress of urban construction, improving the overall image and quality of the city, and facilitating the traffic travel of private car residents. With the increase in the number of residents traveling by car, some problems have emerged in the actual operation of traffic lights. Since the timing schemes of each signal on many traffic lights are relatively fixed and cannot be adjusted in a timely manner according to the actual traffic flow changes in different periods, road congestion occurs in urban traffic. During the morning and evening rush hours, the traffic flow increases greatly, but the signal timing may be the same as that in the off-peak period, resulting in increased congestion at intersections. In the low-peak stage, due to the lack of corresponding adjustment of the signal timing, the waiting time of vehicles may be too long, and at the same time, there may be a situation of empty waiting at some intersections. Therefore, according to the actual situation of the intersection, reasonably and specifically optimizing the timing of each signal on the traffic light can provide a more convenient traffic environment for residents traveling by car, help improve the traffic efficiency at intersections, relieve traffic congestion, and enhance the happiness of residents' travel. Summary of the Invention

[0003] The purpose of the present invention is to provide an intelligent urban planning management system and method based on big data to solve the problems raised in the prior art.

[0004] To achieve the above purpose, the present invention provides the following technical solutions:

[0005] An intelligent urban planning management method based on big data, comprising the following steps:

[0006] Step S100: Obtain the plane layout diagram of the target intersection. The plane layout diagram includes several lanes connecting the target intersection. Obtain the historical surveillance video of the target intersection, and according to the target turning behaviors of the vehicles on each lane after passing through the target intersection, obtain the turning proportion vectors corresponding to each lane;

[0007] Step S200: Combine several lanes into a lane group, extract the target video segments therefrom according to the surveillance video, and calculate the current congestion degree of each lane group according to the staying time of each vehicle at the target intersection;

[0008] Step S300: Obtain the effective signals that allow the vehicles on each lane to pass through the target intersection. According to the target turning behaviors and turning proportion vectors of the vehicles on each lane, obtain the characteristic staying times corresponding to each lane, and obtain the timing coefficients of the effective signals corresponding to each lane according to the congestion degree;

[0009] For example, at an intersection, when the green light of the east-west straight lane lights up, this signal is the valid signal for this lane. At this time, vehicles going straight in the east-west direction can pass through the intersection.

[0010] Step S400: Obtain the historical timing data of the valid signals corresponding to each lane, and adjust the timing data of the valid signals corresponding to each lane according to the timing coefficient.

[0011] Furthermore, step S100 includes:

[0012] Step S110: Obtain the surveillance videos within the previous T days, obtain the target intersection region TIR and each lane region in the surveillance videos according to the floor plan; obtain the rectangular bounding box corresponding to a certain vehicle in the surveillance video as the vehicle region CR of the certain vehicle, and obtain the time period when TIR ∩ CR is not empty. Take the smallest moment in the time period as M1 and the largest moment as M2; take the position of the center point of the vehicle region CR as the midpoint position, set the time period TL0, take the midpoint position at the moment TL0 before the smallest moment M1 as the starting point of a certain vehicle as PT1, the midpoint position at the smallest moment M1 as P1, and the direction from the midpoint position PT1 to the midpoint position P1 as the moving direction D1; take the midpoint position at the largest moment M2 of a certain vehicle as P2, take the midpoint position at the moment TL0 after the largest moment M2 as the starting point as PT2, and the direction from the midpoint position P2 to the midpoint position PT2 as the moving direction D2. The target turning behaviors include going straight, turning right, turning left, and U-turn; determine the to-be-determined turning behavior of a certain vehicle according to the angle required to rotate the moving direction D1 clockwise to the moving direction D2.

[0013] Among them, the rectangular bounding box corresponding to a certain vehicle can be obtained through object detection algorithms such as Faster R-CNN and YOLO; these algorithm models can identify vehicles in images through training with a large amount of image data. When the model outputs the position of the vehicle in the image, it is represented in the form of a rectangular bounding box. For surveillance videos, object detection can be continuously performed on each frame of the image to track the position and movement of the vehicle in real time.

[0014] Step S120: Determine the set of vehicle allowed turning behaviors corresponding to each lane according to traffic rules; obtain the vehicle region CR1 of a certain vehicle at the smallest moment M1, obtain the intersection area between the vehicle region CR1 and a certain lane region. If the intersection area is greater than the lane intersection area threshold, obtain the lane Lan corresponding to the certain lane region, and take the certain vehicle as the target vehicle on the lane Lan. If the set of vehicle allowed turning behaviors corresponding to the lane Lan includes the to-be-determined turning behavior of the certain vehicle, then take the to-be-determined turning behavior as the target turning behavior of the certain vehicle.

[0015] When the vehicle is turning, if most of the vehicle is in a certain lane, it is possible to more accurately determine from which lane the vehicle is turning, facilitating relevant calculations and analyses of the vehicles in that lane.

[0016] Step S130: Obtain the target vehicles corresponding to each lane in the surveillance video. Take the total number of target vehicles corresponding to a certain lane as SUM0. According to the target turning behaviors of the target vehicles on a certain lane, gather the number of vehicles corresponding to the same target turning behavior to obtain the total number of vehicles corresponding to each target turning behavior. Based on the total number of target vehicles SUM0, obtain the turning ratios corresponding to each target turning behavior on a certain lane, and establish a turning ratio vector V0 corresponding to a certain lane.

[0017] Due to the different road conditions at each intersection, when the left-turn indicator light is green, the vehicle may either make a U-turn or turn left. In this solution, the turning ratio vector V0 corresponding to a certain lane = [R1, R2, R3, R4], where R1, R2, R3, and R4 respectively represent the proportions of the straight, right-turn, left-turn, and U-turn directions of the vehicles on a certain lane, and R1, R2, R3, and R4 are all between 0 and 1, and R1 + R2 + R3 + R4 = 1.

[0018] Further, step S200 includes:

[0019] Step S210: Take all adjacent lanes with the same permitted vehicle turning behavior as a lane group, and obtain the lane group area in the surveillance video; obtain the effective signal that permits the vehicles on the lane group to pass through the target intersection, and obtain the total effective signal duration TP corresponding to the effective signal; intercept the surveillance video segment within the effective signal, and obtain the target video segment that is closest to the current time and whose duration is equal to the total effective signal duration TP.

[0020] Step S220: Take the vehicle area corresponding to a certain vehicle at time m in the target video segment as CR0, obtain the intersection area between the vehicle area CR0 and the lane group area. If the intersection area is greater than a pre-set lane group intersection area threshold and at time m + 1, TIR ∩ CR0 is not empty, where TIR is the target intersection area, then take a certain vehicle as a characteristic vehicle, and take the lane group corresponding to the lane group area as Lag; further obtain the total number of characteristic vehicles corresponding to the lane group Lag in the target video segment, and sort the characteristic vehicles in the order from front to back according to the time when each characteristic vehicle is first TIR ∩ CR1 is not empty; and take the total duration when a certain characteristic vehicle satisfies TIR ∩ CR1 is not empty as the residence duration of the certain characteristic vehicle in the target intersection area.

[0021] Step S230: Obtain any two adjacent characteristic vehicles a and b. The characteristic vehicles a and b are M a and M b respectively at the moment when the first TIR ∩ CR1 is not empty. Take the absolute value of the difference between the moments M a and M b as the time difference, obtain all the time differences corresponding to a certain lane group, and obtain the current congestion level of a certain lane group according to the residence duration corresponding to each characteristic vehicle , and perform normalization. Among them, K RT is the residence duration coefficient, K MD is the time difference coefficient, G is the total number of characteristic vehicles corresponding to a certain lane group, H is the total number of time differences corresponding to a certain lane group, G = H + 1, RT g is the residence duration corresponding to the gth characteristic vehicle, and MD h is the hth time difference.

[0022] If the residence time shows a decreasing trend, and the time difference between two adjacent characteristic vehicles shows an increasing trend, it is considered that there are fewer vehicles on the lane at this time, and it is less crowded; the congestion level ranges from [0, 1]. When the congestion level value is larger, it means that there are more vehicles on the lane; in this solution, the congestion level is calculated based on the residence time of characteristic vehicles within a period of time and the time difference between two adjacent characteristic vehicles. This is because the dynamic vehicle data changes at the intersection can better reflect the congestion level of the lane;

[0023] Furthermore, step S300 includes:

[0024] Step S310: According to the target turning behavior of each vehicle on a certain lane within the target intersection area, collect the number of vehicles corresponding to the same target turning behavior. The total number of vehicles corresponding to a certain target turning behavior is several, and according to the residence duration of each vehicle, obtain the average residence duration corresponding to a certain target turning behavior, and then obtain the residence duration vector. Take the dot product of the turning ratio vector corresponding to a certain lane and the residence duration vector as the characteristic residence duration corresponding to a certain lane;

[0025] The residence duration vector is obtained based on the residence time of the vehicle at the target intersection. The residence duration vector S0 corresponding to a certain lane = [s1, s2, s3, s4]; for example, if on a certain lane, only left turns or U-turns are allowed at the target intersection, and the corresponding turning ratio vector V0 of this lane = [0, 0, 0.6, 0.4], and the residence duration vector S0 = [0, 0, 8, 10], then the characteristic residence duration corresponding to a certain lane is the dot product of V0 and S0, which is 0*0 + 0*0 + 0.6*8 + 0.4*10 = 8.8, with the unit of seconds;

[0026] Step S320: If a certain lane is included in a certain lane group, then take the congestion level of the certain lane group as the congestion level of the certain lane, and based on the characteristic residence duration corresponding to the certain lane, obtain the timing coefficient of the effective signal corresponding to each lane as C = k * Y * Z, and perform normalization, where k is the timing correlation coefficient, Y is the congestion level corresponding to the certain lane, and Z is the characteristic residence duration corresponding to the certain lane.

[0027] The timing coefficient indicates the correlation coefficient for the effective signal corresponding to the traffic signal that needs to be re-timed. The value range of the timing coefficient is [0, 1]. When it is close to 0, it means that the traffic flow on the lane in that direction is relatively small, or the currently allocated green light time is too long, resulting in a waste of time, and the current corresponding effective signal needs to reduce the number of seconds. When it is close to 1, it means that the traffic flow on the lane in that direction is relatively large, or the currently allocated green light time is too short, resulting in traffic congestion, indicating that the corresponding effective signal needs to increase the number of seconds.

[0028] Further, step S400 includes:

[0029] Step S410: Obtain the effective signal corresponding to each lane. According to the total effective signal duration of each effective signal, intercept all the monitoring video segments corresponding to each effective signal; take the total effective signal duration of a certain effective signal corresponding to a certain lane as SD. According to step S200, obtain the congestion level corresponding to the certain lane in each monitoring video segment; mark the monitoring video segments with a congestion level less than the first congestion level for the first time, and based on the maximum value W1 of the total number of characteristic vehicles corresponding to the monitoring video segments marked for the first time, obtain the minimum signal duration T1 = SD * (1 - ), where K1 is the first signal coefficient and K1 > 0; mark the monitoring video segments with a congestion level greater than the second congestion level for the second time, and based on the minimum value W2 of the total number of characteristic vehicles corresponding to the monitoring video segments marked for the second time, and the characteristic residence duration S of the certain lane, obtain the maximum signal duration T2 = SD * , where K2 is the second signal coefficient and K2 > 0;

[0030] Among them, the value of the minimum signal duration T1 should be as large as possible under the condition of being less than the total effective signal duration SD, because the duration after re-timing should not differ greatly from the total effective signal duration SD; in y = 1 - e -xAmong them, when the value range of x is x≥0, the value range of y is 1>y≥0, and y increases as x increases; so in this solution, the larger W1 is, the larger T1 obtained is. Therefore, it is necessary to obtain the maximum value W1 of the total number of characteristic vehicles corresponding to the monitored video segment of the first marker, and the first signal coefficient K1 is determined according to the actual situation; the value of the maximum signal duration T2 should be as small as possible while satisfying being greater than the total effective signal duration SD, because the duration after re-timing should not differ greatly from the total effective signal duration SD; when y = e x Among them, when the value range of x is x≥0, the value range of y is y≥1, and y decreases as x decreases; so in this solution, the smaller W2 is, the smaller T2 obtained is. Therefore, it is necessary to obtain the minimum value W2 of the total number of characteristic vehicles corresponding to the monitored video segment of the second marker, and the second signal coefficient K2 is determined according to the actual situation. In this embodiment, in order to make the obtained data more reasonable, the coefficients K1 and K2 should be adjusted so that the values of T1 and T2 are within a reasonable range;

[0031] Step S420: According to the total effective signal duration SD of a certain effective signal and the timing coefficient C, the re-timed signal duration of a certain effective signal is obtained as T0 = SD(1 + K SD *(C - 0.5)), where K SD is the characteristic timing coefficient. When T0 < T1, let T0 = T1; when T0 > T2, let T0 = T2.

[0032] An intelligent urban planning and management system based on big data includes a steering ratio vector obtaining module, a congestion degree calculation module, a timing coefficient calculation module, and an effective signal adjustment module;

[0033] Steering ratio vector obtaining module: used to obtain the planar layout diagram of the target intersection. The planar layout diagram includes several lanes connecting the target intersection, obtain the historical monitored video of the target intersection, and obtain the steering ratio vector corresponding to each lane according to the target steering behavior of the vehicles on each lane after passing through the target intersection;

[0034] Congestion degree calculation module: used to combine several lanes into a lane group, extract the target video segment from the monitored video according to the monitored video, and calculate the current congestion degree of each lane group according to the residence time of each vehicle at the target intersection;

[0035] Timing coefficient calculation module: used to obtain the effective signal that allows the vehicles on each lane to pass through the target intersection corresponding to each lane, obtain the characteristic residence time corresponding to each lane according to the target steering behavior and steering ratio vector of the vehicles on each lane, and obtain the timing coefficient of the effective signal corresponding to each lane according to the congestion degree;

[0036] Adjust the effective signal module: used to obtain the historical timing data of the effective signals corresponding to each lane, and adjust the timing data of the effective signals corresponding to each lane according to the timing coefficient.

[0037] Furthermore, the steering ratio vector obtaining module includes a to-be-determined steering behavior determination unit, a target steering behavior obtaining unit, and a steering ratio vector calculation unit;

[0038] To-be-determined steering behavior determination unit: used to obtain the monitoring videos within the previous T days, obtain the target intersection area and each lane area in the monitoring videos according to the plane layout diagram; the target steering behaviors include going straight, turning right, turning left, and U-turn, and determine the to-be-determined steering behavior of a certain vehicle.

[0039] Target steering behavior obtaining unit: used to determine the set of allowable steering behaviors of vehicles corresponding to each lane according to traffic rules, obtain the vehicle area of a certain vehicle at the minimum moment, and further obtain the target steering behavior of a certain vehicle.

[0040] Steering ratio vector calculation unit: used to obtain the target vehicles corresponding to each lane in the monitoring video, and obtain the total number of target vehicles corresponding to a certain lane; and obtain the total number of vehicles corresponding to each target steering behavior, obtain the steering ratios corresponding to each target steering behavior on a certain lane, and establish a steering ratio vector corresponding to a certain lane.

[0041] Furthermore, the timing coefficient calculation module includes a characteristic residence duration calculation unit and a timing coefficient calculation unit;

[0042] Characteristic residence duration calculation unit: used to obtain the average residence duration corresponding to a certain target steering behavior according to the residence duration of each vehicle, and further obtain a residence duration vector, and take the dot product of the steering ratio vector corresponding to a certain lane and the residence duration vector as the characteristic residence duration corresponding to a certain lane;

[0043] Timing coefficient calculation unit: used to regard the congestion degree of a certain lane group as the congestion degree of a certain lane, and obtain the timing coefficient of the effective signal corresponding to each lane according to the characteristic residence duration corresponding to a certain lane.

[0044] Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention provides an intelligent urban planning and management system and method based on big data, including: obtaining a plane layout diagram of a target intersection and historical surveillance videos, and obtaining a turning ratio vector corresponding to each lane; extracting target video segments therefrom according to the surveillance videos, and then calculating the current congestion degree of each lane group; obtaining the effective signals that allow vehicles on each lane to pass through the target intersection, and obtaining the timing coefficients of the effective signals corresponding to each lane; obtaining the historical timing data of the effective signals corresponding to each lane, and adjusting the timing data of the effective signals corresponding to each lane according to the timing coefficients. By obtaining the historical surveillance data of the target intersection, and further calculating the current congestion degree of the lane group according to the actual situation of the intersection, the present invention rationally and specifically optimizes the timing of each signal on the traffic lights, provides a more convenient traffic environment for residents driving, improves the traffic efficiency of the intersection, alleviates traffic congestion, and enhances the happiness of residents' travel. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a schematic flow chart of an intelligent urban planning and management method based on big data according to the present invention;

[0046] Figure 2 It is a structural diagram of an intelligent urban planning and management system based on big data according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0047] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0048] Embodiment: As Figure 1 shown, the present invention provides a technical solution for an intelligent urban planning and management method based on big data, including the following steps:

[0049] Step S100: Obtain a plane layout diagram of a target intersection, where the plane layout diagram includes several lanes connecting the target intersection, obtain historical surveillance videos of the target intersection, and obtain a turning ratio vector corresponding to each lane according to the target turning behaviors of the vehicles on each lane after passing through the target intersection.

[0050] Step S110: Obtain the surveillance videos within the previous T days. Based on the floor plan, obtain the target intersection region TIR and each lane region in the surveillance videos; obtain the rectangular bounding box corresponding to a certain vehicle in the surveillance video as the vehicle region CR of the certain vehicle, and obtain the time period when TIR ∩ CR is not empty. Take the earliest moment in the time period as M1 and the latest moment as M2; take the position where the center point of the vehicle region CR is located as the midpoint position, set the time period TL0, take the midpoint position at the moment TL0 before the earliest moment M1 of the certain vehicle as PT1, and the midpoint position at the earliest moment M1 as P1. Take the direction from the midpoint position PT1 to the midpoint position P1 as the moving direction D1; take the midpoint position at the latest moment M2 of the certain vehicle as P2, take the midpoint position at the moment TL0 after the latest moment M2 as PT2, and take the direction from the midpoint position P2 to the midpoint position PT2 as the moving direction D2. The target turning behaviors include going straight, turning right, turning left, and U-turn; determine the to-be-determined turning behavior of the certain vehicle according to the angle required to rotate the moving direction D1 clockwise to the moving direction D2.

[0051] Among them, the rectangular bounding box corresponding to a certain vehicle can be obtained through object detection algorithms such as Faster R-CNN and YOLO; these algorithm models can identify the vehicles in the images through a large amount of image data training. When the model outputs the position of the vehicle in the image, it is represented in the form of a rectangular bounding box. For the surveillance video, object detection can be continuously performed on each frame of the image to track the position and movement of the vehicle in real time. According to the angle required to rotate the moving direction D1 clockwise to the moving direction D2, the value range of this angle is [0, 360]. In this embodiment, if the angle is 0 - 45 or 315 - 360, it is determined that the to-be-determined turning behavior of the vehicle is a U-turn; if the angle is 45 - 135, it is determined that the to-be-determined turning behavior of the vehicle is a right turn; if the angle is 135 - 225, it is determined that the to-be-determined turning behavior of the vehicle is going straight; if the angle is 225 - 315, it is determined that the to-be-determined turning behavior of the vehicle is a left turn.

[0052] Step S120: Determine the set of allowed turning behaviors for the vehicles corresponding to each lane according to the traffic rules; obtain the vehicle region CR1 of a certain vehicle at the earliest moment M1, obtain the intersection area between the vehicle region CR1 and a certain lane region. If the intersection area is greater than the lane intersection area threshold, obtain the lane Lan corresponding to the certain lane region, and take the certain vehicle as the target vehicle on the lane Lan. If the set of allowed turning behaviors for the lane Lan includes the to-be-determined turning behavior of the certain vehicle, then take the to-be-determined turning behavior as the target turning behavior of the certain vehicle.

[0053] In this embodiment, the lane intersection area threshold is 0.8 * CR1 of the vehicle area. When the vehicle turns, if most of the vehicle area is in a certain lane, it is possible to more accurately determine from which lane the vehicle turns, facilitating relevant calculations and analyses of the vehicles on that lane.

[0054] Step S130: Obtain the target vehicles corresponding to each lane in the surveillance video. Take the total number of target vehicles corresponding to a certain lane as SUM0. According to the target turning behaviors of the target vehicles on a certain lane, aggregate the number of vehicles corresponding to the same target turning behavior to obtain the total number of vehicles corresponding to each target turning behavior. Based on the total number of target vehicles SUM0, obtain the turning ratios corresponding to each target turning behavior on a certain lane, and establish a turning ratio vector V0 corresponding to a certain lane.

[0055] Due to the different road conditions at each intersection, when the left-turn indicator light is green, the vehicle may turn around or turn left. In this solution, the turning ratio vector V0 corresponding to a certain lane = [R1, R2, R3, R4], where R1, R2, R3, and R4 respectively represent the proportions of the straight, right-turn, left-turn, and U-turn directions of the vehicles on a certain lane, and R1, R2, R3, and R4 are all between 0 and 1, and R1 + R2 + R3 + R4 = 1. For example, if the turning ratios corresponding to each target turning behavior on a certain lane are [0, 0, 0.6, 0.4], then the turning ratio vector V0 = [0, 0, 0.6, 0.4].

[0056] Step S200: Combine several lanes into a lane group. Extract the target video segment therefrom according to the surveillance video, and calculate the current congestion degree of each lane group based on the residence time of each vehicle at the target intersection.

[0057] Step S210: Take all adjacent lanes with the same permitted vehicle turning behavior as a lane group, and obtain the lane group area in the surveillance video; obtain the valid signal that permits the vehicles on the lane group to pass through the target intersection, and obtain the total duration TP of the valid signal corresponding to the valid signal; intercept the surveillance video segment within the valid signal, and obtain the target video segment that is the closest to the current time and whose duration is equal to the total duration TP of the surveillance video segment.

[0058] Step S220: Take the vehicle area corresponding to a certain vehicle in the target video segment at time m as CR0, obtain the intersection area between the vehicle area CR0 and a certain lane group area. If the intersection area is greater than the pre-set lane group intersection area threshold and TIR ∩ CR0 is not empty at time m + 1, where TIR is the target intersection area, then take the certain vehicle as a characteristic vehicle, and the lane group corresponding to the certain lane group area as Lag; further obtain the total number of characteristic vehicles corresponding to the lane group Lag in the target video segment, and sort the characteristic vehicles in the order from front to back according to the time when each characteristic vehicle is first non-empty for TIR ∩ CR1; and take the total duration when a certain characteristic vehicle satisfies TIR ∩ CR1 is not empty as the residence duration of the certain characteristic vehicle in the target intersection area.

[0059] Step S230: Obtain any two adjacent characteristic vehicles a and b in sequence numbers. The times when the characteristic vehicles a and b are first non-empty for TIR ∩ CR1 are M a and M b , respectively. Take the absolute value of the difference between the times M a and M b as the time difference, obtain all the time differences corresponding to a certain lane group, and obtain the current congestion degree of a certain lane group according to the residence duration corresponding to each characteristic vehicle. , and perform normalization. Among them, K RT is the residence duration coefficient, K MD is the time difference coefficient, G is the total number of characteristic vehicles corresponding to a certain lane group, H is the total number of time differences corresponding to a certain lane group, G = H + 1, RT g is the residence duration corresponding to the g-th characteristic vehicle, MD h is the h-th time difference.

[0060] If the residence time shows a decreasing trend, and the time difference between two adjacent characteristic vehicles in sequence numbers shows an increasing trend, it is considered that there are fewer vehicles on the lane at this time, and it is less crowded; the congestion degree ranges from [0, 1]. The larger the congestion degree value, the more vehicles there are on the lane; in this solution, the congestion degree is calculated based on the residence time of characteristic vehicles within a period of time and the time difference between two adjacent characteristic vehicles in sequence numbers, because the dynamic vehicle data changes at the intersection can better reflect the congestion degree of the lane.

[0061] Step S300: Obtain the effective signal that allows the vehicles on each lane to pass through the target intersection, obtain the characteristic residence duration corresponding to each lane according to the target turning behaviors and turning ratio vectors of the vehicles on each lane, and obtain the timing coefficient of the effective signal corresponding to each lane according to the congestion degree.

[0062] For example, at an intersection, when the green light for the east-west straight-through lane lights up, this signal is the valid signal for that lane, and at this time, vehicles going straight east-west can pass through the intersection.

[0063] Step S310: According to the target turning behaviors of each vehicle in a certain lane within the target intersection area, gather the number of vehicles corresponding to the same target turning behavior. The total number of several vehicles corresponds to a certain target turning behavior, and based on the residence duration of each vehicle, obtain the average residence duration corresponding to a certain target turning behavior, and then obtain the residence duration vector. Take the dot product of the turning ratio vector corresponding to a certain lane and the residence duration vector as the characteristic residence duration corresponding to a certain lane.

[0064] The residence duration vector is obtained based on the residence time of the vehicle at the target intersection. The residence duration vector S0 corresponding to a certain lane = [s1, s2, s3, s4]; for example, if only left turns or U-turns are allowed in a certain lane at the target intersection, and the corresponding turning ratio vector V0 of this lane = [0, 0, 0.6, 0.4], and the residence duration vector S0 = [0, 0, 8, 10], then the characteristic residence duration corresponding to a certain lane is the dot product of V0 and S0, which is 0*0 + 0*0 + 0.6*8 + 0.4*10 = 8.8, with the unit of seconds.

[0065] Step S320: If a certain lane is included in a certain lane group, take the congestion degree of the certain lane group as the congestion degree of the certain lane, and based on the characteristic residence duration corresponding to a certain lane, obtain the timing coefficient C = k*Y*Z for the valid signal corresponding to each lane, and perform normalization, where k is the timing correlation coefficient, Y is the congestion degree corresponding to a certain lane, and Z is the characteristic residence duration corresponding to a certain lane.

[0066] The timing coefficient indicates the correlation coefficient for the re-timing required for the valid signal corresponding to the traffic signal. The value range of the timing coefficient is [0, 1]. When it is close to 0, it means that the traffic flow on the lane in that direction is relatively small, or the currently allocated green light time is too long, resulting in a waste of time, and the currently corresponding valid signal needs to reduce the number of seconds. When it is close to 1, it means that the traffic flow on the lane in that direction is relatively large, or the currently allocated green light time is too short, resulting in traffic congestion, indicating that the corresponding valid signal needs to increase the number of seconds.

[0067] Step S400: Obtain the historical timing data of the valid signal corresponding to each lane, and adjust the timing data of the valid signal corresponding to each lane according to the timing coefficient.

[0068] Step S410: Obtain the valid signals corresponding to each lane. According to the total duration of the valid signals of each valid signal, intercept all the monitored video segments corresponding to each valid signal; use the total duration of the valid signal of a certain valid signal corresponding to a certain lane as SD. According to step S200, obtain the congestion level corresponding to each lane in each monitored video segment; mark the monitored video segments with a congestion level less than the first congestion level, and based on the maximum value W1 of the total number of characteristic vehicles corresponding to the monitored video segments marked with the first mark, obtain the minimum signal duration T1 = SD * (1 - ), where K1 is the first signal coefficient and K1 > 0; mark the monitored video segments with a congestion level greater than the second congestion level, and based on the minimum value W2 of the total number of characteristic vehicles corresponding to the monitored video segments marked with the second mark, and the characteristic residence duration S of a certain lane, obtain the maximum signal duration T2 = SD * , where K2 is the second signal coefficient and K2 > 0.

[0069] Among them, the value of the minimum signal duration T1 should be as large as possible under the condition of being less than the total duration SD of the valid signal, because the duration after re-timing should not differ greatly from the total duration SD of the valid signal; in y = 1 - e -x , when the value range of x is x ≥ 0, the value range of y is 1 > y ≥ 0, and y increases as x increases; so in this solution, when W1 is larger, the obtained T1 is larger, so it is necessary to obtain the maximum value W1 of the total number of characteristic vehicles corresponding to the monitored video segments marked with the first mark. The first signal coefficient K1 is determined according to the actual situation. In this embodiment, K1 = 0.1; the value of the maximum signal duration T2 should be as small as possible under the condition of being greater than the total duration SD of the valid signal, because the duration after re-timing should not differ greatly from the total duration SD of the valid signal; in y = e x , when the value range of x is x ≥ 0, the value range of y is y ≥ 1, and y decreases as x decreases; so in this solution, when W2 is smaller, the obtained T2 is smaller, so it is necessary to obtain the minimum value W2 of the total number of characteristic vehicles corresponding to the monitored video segments marked with the second mark. The second signal coefficient K2 is determined according to the actual situation. In this embodiment, K2 = 0.01. In this embodiment, in order to make the obtained data more reasonable, the coefficients K1 and K2 should be adjusted so that the values of T1 and T2 are within a reasonable range. Generally, the value of T1 is not easily less than 0.8 * SD, and the value of T1 is not easily greater than 1.2 * SD.

[0070] Step S420: According to the total duration SD of a certain valid signal and the timing coefficient C, obtain the signal duration after re-timing of a certain valid signal as T0 = SD(1 + K SD * (C - 0.5)), K SDIt is the characteristic timing coefficient. When T0 < T1, let T0 = T1. When T0 > T2, let T0 = T2.

[0071] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above-described exemplary embodiments, and the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be embraced within the present invention. Any reference signs in the claims should not be construed as limiting the claims involved.

Claims

1. An intelligent urban planning and management method based on big data, characterized in that, Including the following steps: Step S100: Obtain the planar layout diagram of the target intersection. The planar layout diagram includes several lanes connecting the target intersection. Obtain the historical surveillance video of the target intersection. According to the target turning behaviors of the vehicles on each lane after passing through the target intersection, obtain the turning ratio vector corresponding to each lane; Step S200: Combine several lanes into a lane group. According to the surveillance video, extract the target video segment therein, and calculate the current congestion degree of each lane group according to the residence time of each vehicle at the target intersection; Step S300: Obtain the effective signal that allows the vehicles on each lane to pass through the target intersection. According to the target turning behaviors and turning ratio vectors of the vehicles on each lane, obtain the characteristic residence time corresponding to each lane, and obtain the timing coefficient of the effective signal corresponding to each lane according to the congestion degree; Step S400: Obtain the historical timing data of the effective signal corresponding to each lane, and adjust the timing data of the effective signal corresponding to each lane according to the timing coefficient.

2. The intelligent city planning and management method based on big data according to claim 1, characterized in that, Step S100 includes: Step S110: Obtain the surveillance video within the previous T days. According to the planar layout diagram, obtain the target intersection region TIR and each lane region in the surveillance video; Obtain the rectangular bounding box corresponding to a certain vehicle in the surveillance video as the vehicle region CR of the certain vehicle, and obtain the time period when TIR ∩ CR is not empty. Take the smallest moment in the time period as M1 and the largest moment as M2; Take the position of the center point of the vehicle region CR as the midpoint position, set the time period TL0, take the midpoint position at the moment TL0 before the smallest moment M1 as the starting point of the certain vehicle as PT1, the midpoint position at the smallest moment M1 as P1, and the direction from the midpoint position PT1 to the midpoint position P1 as the moving direction D1; Take the midpoint position at the largest moment M2 of the certain vehicle as P2, take the midpoint position at the moment TL0 after the largest moment M2 as the starting point as PT2, and the direction from the midpoint position P2 to the midpoint position PT2 as the moving direction D2. The target turning behaviors include going straight, turning right, turning left, and U-turn; Determine the to-be-determined turning behavior of the certain vehicle according to the angle required for the moving direction D1 to rotate clockwise to the moving direction D2; Step S120: Determine the set of vehicle allowed turning behaviors corresponding to each lane according to traffic rules; Obtain the vehicle region CR1 of the certain vehicle at the smallest moment M1, obtain the intersection area between the vehicle region CR1 and a certain lane region. If the intersection area is greater than the lane intersection area threshold, obtain the lane Lan corresponding to the certain lane region, and take the certain vehicle as the target vehicle on the lane Lan. If the set of vehicle allowed turning behaviors corresponding to the lane Lan includes the to-be-determined turning behavior of the certain vehicle, then take the to-be-determined turning behavior as the target turning behavior of the certain vehicle; Step S130: Obtain the target vehicles corresponding to each lane in the monitoring video, take the total number of target vehicles corresponding to a certain lane as SUM0, and according to the target turning behaviors of the target vehicles on a certain lane, collect the number of vehicles corresponding to the same target turning behavior, obtain the total number of vehicles corresponding to each target turning behavior, and according to the total number of target vehicles SUM0, obtain the turning ratio corresponding to each target turning behavior on a certain lane, and establish a turning ratio vector V0 corresponding to a certain lane.

3. The intelligent city planning and management method based on big data according to claim 2, wherein, Step S200 includes: Step S210: Take all adjacent lanes with the same permitted vehicle turning behavior as a lane group, and obtain the lane group area in the monitoring video; obtain the valid signal that permits the vehicles on the lane group to pass through the target intersection corresponding to a certain lane group, and obtain the total duration TP of the valid signal corresponding to the valid signal; intercept the monitoring video segment within the valid signal, and obtain the target video segment that is closest to the current time and the duration of the monitoring video segment is equal to the total duration TP of the valid signal; Step S220: Take the vehicle area corresponding to a certain vehicle at time m in the target video segment as CR0, obtain the intersection area between the vehicle area CR0 and a certain lane group area. If the intersection area is greater than the pre-set lane group intersection area threshold and TIR∩CR0 is not empty at time m + 1, where TIR is the target intersection area, then take the certain vehicle as a characteristic vehicle, and take the lane group corresponding to the certain lane group area as Lag; furthermore, obtain the total number of characteristic vehicles corresponding to the lane group Lag in the target video segment, and sort the characteristic vehicles in the order from front to back according to the time when each characteristic vehicle is first TIR∩CR1 is not empty; and take the total duration when a certain characteristic vehicle satisfies TIR∩CR1 is not empty as the residence duration of the certain characteristic vehicle in the target intersection area. Step S230: Obtain any two adjacent characteristic vehicles a and b. The characteristic vehicles a and b are M a and M b respectively at the moment when the first TIR ∩ CR1 is not empty. Take the absolute value of the difference between the moments M a and M b as the time difference, obtain all the time differences corresponding to the certain lane group, and obtain the current congestion degree of the certain lane group according to the residence duration corresponding to each characteristic vehicle , and perform normalization. Among them, K RT is the residence duration coefficient, K MD is the time difference coefficient, G is the total number of characteristic vehicles corresponding to the certain lane group, H is the total number of time differences corresponding to the certain lane group, G = H + 1, RT g is the residence duration corresponding to the g-th characteristic vehicle, MD h is the h-th time difference.

4. An intelligent urban planning and management method based on big data according to claim 3, characterized in that, Step S300 includes: Step S310: According to the target turning behaviors of each vehicle on a certain lane in the target intersection area, collect the number of vehicles corresponding to the same target turning behavior, the total number of several vehicles corresponding to a certain target turning behavior, and according to the residence duration of each vehicle, obtain the average residence duration corresponding to a certain target turning behavior, and then obtain a residence duration vector, and take the dot product of the turning ratio vector corresponding to a certain lane and the residence duration vector as the characteristic residence duration corresponding to a certain lane. Step S320: If a certain lane is included in a certain lane group, take the congestion degree of the certain lane group as the congestion degree of the certain lane, and according to the characteristic residence duration corresponding to a certain lane, obtain the timing coefficient C = k*Y*Z of the valid signal corresponding to each lane, and perform normalization, where k is a timing-related coefficient, Y is the congestion degree corresponding to a certain lane, and Z is the characteristic residence duration corresponding to a certain lane.

5. The intelligent city planning and management method based on big data according to claim 4, characterized in that, Step S400 includes: Step S410: Obtain the valid signals corresponding to each lane. According to the total duration of the valid signals of each valid signal, intercept all the monitored video segments corresponding to each valid signal; use the total duration of the valid signal of a certain valid signal corresponding to a certain lane as SD. According to Step S200, obtain the congestion level corresponding to each monitored video segment of a certain lane; mark the monitored video segments with a congestion level less than the first congestion level, and based on the maximum value W1 of the total number of characteristic vehicles corresponding to the monitored video segments marked with the first mark, obtain the minimum signal duration T1 = SD * (1 - ), where K1 is the first signal coefficient and K1 > 0; mark the monitored video segments with a congestion level greater than the second congestion level, and based on the minimum value W2 of the total number of characteristic vehicles corresponding to the monitored video segments marked with the second mark, and the characteristic residence duration S of the certain lane, obtain the maximum signal duration T2 = SD * , where K2 is the second signal coefficient and K2 > 0; Step S420: According to the total duration SD of a certain valid signal and the timing coefficient C, the signal duration after re-timing of the certain valid signal is obtained as T0 = SD(1 + K SD *(C - 0.5)), where K SD is the characteristic timing coefficient. When T0 < T1, let T0 = T1; when T0 > T2, let T0 = T2.

6. An intelligent urban planning and management system for implementing an intelligent urban planning and management method based on big data according to any one of claims 1-5, characterized in that, The system includes a turning ratio vector obtaining module, a congestion degree calculation module, a timing coefficient calculation module, and an effective signal adjustment module; Steering Ratio Vector Obtaining Module: It is used to obtain the planar layout diagram of the target intersection. The planar layout diagram includes several lanes connecting the target intersection. Obtain the historical surveillance video of the target intersection, and based on the target steering behaviors of the vehicles on each lane after passing through the target intersection, obtain the steering ratio vector corresponding to each lane; Congestion Degree Calculation Module: It is used to combine several lanes into a lane group, extract the target video segment from the surveillance video according to the surveillance video, and calculate the current congestion degree of each lane group based on the staying time of each vehicle at the target intersection; Timing Coefficient Calculation Module: It is used to obtain the effective signal that allows the vehicles on each lane to pass through the target intersection, obtain the characteristic staying time corresponding to each lane based on the target steering behaviors and steering ratio vectors of the vehicles on each lane, and obtain the timing coefficient of the effective signal corresponding to each lane according to the congestion degree; Effective Signal Adjustment Module: It is used to obtain the historical timing data of the effective signal corresponding to each lane, and adjust the timing data of the effective signal corresponding to each lane according to the timing coefficient.

7. An intelligent urban planning and management system according to claim 6, characterized in that, The Steering Ratio Vector Obtaining Module includes a To-be-determined Steering Behavior Obtaining Unit, a Target Steering Behavior Obtaining Unit, and a Steering Ratio Vector Calculation Unit; To-be-determined Steering Behavior Obtaining Unit: It is used to obtain the surveillance video within the previous T days, obtain the target intersection area and each lane area in the surveillance video according to the planar layout diagram; The target steering behaviors include going straight, turning right, turning left, and U-turn. Determine the to-be-determined steering behavior of a certain vehicle; Target Steering Behavior Obtaining Unit: It is used to determine the set of allowable steering behaviors of the vehicles corresponding to each lane according to traffic rules, obtain the vehicle area of a certain vehicle at the minimum moment, and further obtain the target steering behavior of a certain vehicle; Steering Ratio Vector Calculation Unit: It is used to obtain the target vehicles corresponding to each lane in the surveillance video, and obtain the total number of target vehicles corresponding to a certain lane; And obtain the total number of vehicles corresponding to each target steering behavior, obtain the steering ratio corresponding to each target steering behavior on a certain lane, and establish the steering ratio vector corresponding to a certain lane.

8. An intelligent urban planning and management system according to claim 7, characterized in that, The Timing Coefficient Calculation Module includes a Characteristic Staying Time Calculation Unit and a Timing Coefficient Calculation Unit; Characteristic Staying Time Calculation Unit: It is used to obtain the average staying time corresponding to a certain target steering behavior according to the staying time of each vehicle, and further obtain the staying time vector, and use the dot product of the steering ratio vector corresponding to a certain lane and the staying time vector as the characteristic staying time corresponding to a certain lane; Timing Coefficient Calculation Unit: It is used to regard the congestion degree of a certain lane group as the congestion degree of a certain lane, and obtain the timing coefficient of the effective signal corresponding to each lane according to the characteristic staying time corresponding to a certain lane.

Images