A radar video integrated machine traffic flow data outlier detection method and system

Abstract

A method for detecting abnormal values ​​in traffic data of a radar video all-in-one device includes: (1) obtaining traffic data, signal cycle scheme data, and lane steering function data reported by the radar video all-in-one device, and correlating and fusing the above data; (2) judging data omission and duplicate reporting on the fused raw data, marking the raw data with the number of omissions and duplicate reports not exceeding a threshold as normal data, otherwise marking it as abnormal data; (3) performing an abnormality test on the basis of the completeness of the data reported by the device, recording the data that meets the inspection standard of the device during the inspection period as normal data, and recording the data that does not meet the inspection standard as abnormal data; (4) based on the abnormality test of the raw data reported by the device, marking the periodic indicator data settled by the signal control platform. The present invention also includes a radar video all-in-one device traffic data abnormal value detection system. The present invention can serve as a basic basis for the validity of signal control input data.

Description

Technical Field

[0001] The present invention belongs to the field of traffic signal control and relates to a method for detecting abnormality in traffic data, and in particular to a method for detecting abnormality in traffic flow data reported by a radar video integrated device. Background Art

[0002] Compared to cameras and lidar, millimeter-wave radar is less susceptible to weather and offers reliable and accurate target detection. In recent years, it has emerged as an essential sensor in intelligent transportation systems, gradually becoming a key component. However, there are some challenges. Millimeter-wave radar uses the reflection of electromagnetic waves from a target to detect and determine its position. However, cluttered environments (such as large vehicles, objects, signs, and trees) inevitably lead to false alarms. On roads with low traffic, vehicle speeds and positions generally differ significantly, making it easy to distinguish targets. However, on congested roads or complex intersections, where vehicles are moving slowly, pedestrians and non-motorized vehicles mix, and road signs are present, targets can have similar speeds and be very close together. This poses a significant challenge to radar detection capabilities.

[0003] The radar video fusion all-in-one device is a traffic sensor that combines a camera, millimeter-wave radar, and a high-performance processor. Compared to traditional millimeter-wave radars, it has advantages such as a wide detection range, rich detection targets, and all-weather operation. The radar video all-in-one device can report traffic flow data at a preset frequency, including traffic volume, time occupancy, queue length, vehicle type, vehicle trajectory, and so on. The traffic flow data detected by the radar video all-in-one device is used for urban traffic signal control and is used as data input for signal control in scenarios such as single-point optimization control, arterial coordinated control, and regional coordinated control. It effectively improves the traffic efficiency of the road network, reduces the number of unused green lights, and balances the load. The radar video all-in-one device has been widely and maturely applied in urban traffic signal control and intelligent traffic management.

[0004] However, due to the complex structure of the equipment, the road traffic environment, and extreme weather conditions, radar video all-in-one devices still have problems such as missed and misreported data. Signal optimization effects such as adaptive control require high traffic data, and the data input of signal control settles the raw data reported by the device in units of signal cycles, and the settled data cannot be used to trace problems. Therefore, to address the above issues, it is necessary to design a traffic data anomaly detection method for radar video all-in-one devices to effectively ensure the accuracy and effectiveness of signal control data input, empower signal control effects with data, improve the efficiency of intersection traffic operation, and ensure the safety of intersection traffic operation. Summary of the Invention

[0005] In order to solve the above technical problems, the present invention provides a method for detecting abnormality in traffic data of a radar video all-in-one machine.

[0006] The present invention utilizes the data reporting frequency of the radar video all-in-one machine to detect data integrity; and utilizes the traffic flow distribution characteristics, traffic parameter thresholds, traffic parameter correlations, etc. to detect data accuracy. It can effectively ensure that signal control scenarios such as single-point optimization and trunk line coordination can effectively and accurately monitor input data, thereby improving the signal optimization effect.

[0007] In order to achieve the above object, the present invention provides a method for detecting abnormal values ​​in flow data of a radar video integrated machine, comprising the following steps:

[0008] (1) Data acquisition and fusion: Acquire traffic data, signal cycle plan data, and lane steering function data reported by the radar video integrated device, and correlate and fuse the above data;

[0009] (2) Data integrity check: The fused raw data is checked for missed and repeated reports. If the number of missed and repeated reports of the raw data does not exceed the threshold, it is marked as normal data; otherwise, it is marked as abnormal data.

[0010] (3) Data anomaly test: Anomaly test is performed on the basis of the completeness of the data reported by the equipment, including the headway threshold test, traffic flow parameter consistency test, and traffic flow distribution pattern test. The data that meets the test standards during the test period is recorded as normal data, and the data that does not meet the test standards is recorded as abnormal data;

[0011] (4) Data anomaly marking: Based on the original data reported by the equipment, the periodic indicators settled by the signal control platform are marked.

[0012] Furthermore, in step (1) data acquisition and fusion, lanes, phases, and equipment have been bound according to the existing signal control platform, and only the corresponding data need to be merged into tables.

[0013] Furthermore, in step (2) of the data integrity test, the theoretical number of data reported by the radar video integrated device is calculated based on the device reporting frequency and the phase cycle duration of the signal control scheme, and the theoretical value is compared with the actual value. When the difference is within the threshold, it is marked as normal data, otherwise it is marked as abnormal data. The formula for determining data omission and duplication is:

[0014] |μ i ·t ig -n i |≤N1 (1)

[0015] In the above formula:

[0016] μ i - Frequency of radar video integrated device i data reporting (items / s);

[0017] tig -The maximum green light duration (s) of the phase where the radar video integrated device i monitors the lane;

[0018] n i -The number of data items actually reported by the radar video integrated device i;

[0019] N1-threshold for missed and repeated reporting data.

[0020] Furthermore, the data anomaly test in step (3) depends on the following three aspects:

[0021] ① Headway threshold test: Average headway refers to the average time difference between adjacent vehicles passing a certain section, in seconds per vehicle, and is calculated using the following formula:

[0022] t h =3600 / q (2)

[0023] In the above formula:

[0024] t h -headway (s / veh);

[0025] q-flow rate, the number of vehicles passing through a section during the observation period (veh / h);

[0026] When the flow rate q reaches the saturation flow rate, that is, the maximum number of vehicles passing through a section per unit time, and the average headway is minimized, it is considered saturated headway. The headway threshold should be less than the saturated headway. When the headway between adjacent vehicles is greater than the headway threshold, the headway is considered normal; otherwise, it is considered abnormal. The judgment formula is:

[0027]

[0028] In the above formula:

[0029] - The moment when the kth vehicle passes by lane j monitored by radar and video integrated device i;

[0030] - The moment when the k+1th vehicle passes by the radar video integrated device i in lane j;

[0031] t hmin -Threshold of headway;

[0032] ② Traffic flow parameter consistency test: The radar video integrated device detects vehicles through video, and the radar detects the queue length of the lane at the end of the red light. The queue length reported in the same cycle and the number of vehicles passing through should be consistent. That is, the difference between the queue length converted into the number of vehicles and the actual number of vehicles in the cycle should not exceed the threshold. In this case, the traffic flow parameters are consistent; otherwise, the traffic flow parameters are inconsistent and it is judged as abnormal. The judgment formula is:

[0033]

[0034] In the above formula:

[0035] -The maximum queue length (in meters) of lane j in the cycle reported by radar video integrated device i;

[0036] - Radar and video integrated device i reports data to count the number of vehicles in lane j during the cycle;

[0037] - Average vehicle length (m);

[0038] N2 - queue length and traffic consistency difference threshold;

[0039] ③ Traffic flow distribution pattern inspection: The traffic flow distribution pattern of lanes with the same entrance and the same turn at the road intersection should be consistent. That is, the difference in traffic parameters between lanes with the same entrance and the same turn reported by the equipment should not exceed the threshold. That is, the difference in the number of vehicles passing through and the maximum queue length between multiple lanes with the same turn on the same entrance within a cycle is within the threshold range. It is judged as normal data, otherwise it is judged as abnormal data. The judgment formula is:

[0040]

[0041] In the above formula:

[0042] -The maximum queue length (in meters) of lane j in the cycle reported by radar video integrated device i;

[0043] - The maximum queue length (in meters) of lane j+1 in this cycle, reported by radar video integrated device i, which has the same entrance and turn as lane j;

[0044] - Average vehicle length (m);

[0045] N3-the threshold value for the consistency difference in queue length between the same turning lanes on the same entrance (vehicles);

[0046]

[0047] In the above formula:

[0048] - Radar and video integrated device i reports the number of vehicles in lane j during the cycle;

[0049] - Radar and video integrated device i reports the number of vehicles entering and turning into lane j+1 in the same cycle as lane j;

[0050] N4-Flow consistency difference threshold between the same turning lanes of the same entrance (vehicles);

[0051] Furthermore, the radar video integrated machine traffic data includes date, radar video integrated machine number, lane number, phase duration, phase number, detection target, and queue length.

[0052] Furthermore, the data anomaly marking in step (4) is based on the anomaly inspection results of the original data reported by the device in steps (1), (2), and (3), and marks the periodic indicators of the signal control platform settlement.

[0053] The present invention also includes a radar video integrated machine flow data abnormal value detection system, comprising:

[0054] The data acquisition and fusion module is used to obtain traffic data, signal cycle plan data, and lane steering function data reported by the radar video integrated device, and correlate and fuse the above data;

[0055] The data integrity check module is used to judge the data omission and duplication of the fused original data. If the number of omissions and duplications of the original data does not exceed the threshold, it is marked as normal data; otherwise, it is marked as abnormal data.

[0056] The data anomaly inspection module is used to perform anomaly inspection on the basis of the completeness of the data reported by the equipment. Specifically, it includes the headway threshold inspection, traffic flow parameter consistency inspection, and traffic flow distribution pattern inspection. The data that meets the inspection standards during the inspection period is recorded as normal data, and the data that does not meet the inspection standards is recorded as abnormal data.

[0057] The data anomaly marking module is used to mark the periodic indicators of the signal control platform settlement based on the analysis results of the abnormalities of the original data reported by the equipment.

[0058] The present invention also includes a computer-readable storage medium on which a program is stored. When the program is executed by a processor, a method for detecting abnormal values ​​in flow data of a radar video integrated machine of the present invention is implemented.

[0059] The present invention also includes a computing device, including a memory and a processor, wherein the memory stores executable code, and when the processor executes the executable code, a radar video all-in-one machine traffic data anomaly detection method of the present invention is implemented.

[0060] The beneficial effects of the present invention are: by detecting anomalies in the raw data reported by the radar video integrated device and marking the signal period indicator, the accuracy of signal optimization input data is ensured from the source, improving the effectiveness and accuracy of signal optimization. In particular, it can ensure the security of real-time signal adjustment such as adaptive adjustment, while making the evaluation of signal optimization effects traceable and explainable. The present invention can serve as a foundation for the validity of signal control input data. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 It is a schematic flow diagram of the method of the present invention.

[0062] Figure 2 It is a schematic diagram of the relationship between the equipment, intersection and lane of the present invention.

[0063] Figure 3 This is a schematic diagram of intersection channelization according to one embodiment of the present invention.

[0064] Figure 4 4 is a phase sequence and lane relationship diagram of an embodiment of the present invention.

[0065] Figure 5 It is a schematic diagram of the system structure of the present invention. DETAILED DESCRIPTION

[0066] The following further illustrates the specific implementation of the present invention with reference to the accompanying drawings and examples. However, the implementation and protection of the present invention are not limited thereto. It should be noted that any details not specifically described below can be implemented by those skilled in the art with reference to the prior art. Those skilled in the art should recognize that the present invention encompasses all possible alternatives, improvements, and equivalents within the scope of the claims.

[0067] Example 1

[0068] See also Figure 1 、 2 , 3, 4, take a city's signal control platform that relies on radar video integrated machine traffic data as an example, and perform traffic data outlier detection. A radar video integrated machine traffic data outlier detection method, the steps are as follows:

[0069] (1) Data acquisition and fusion

[0070] Obtain traffic data, signal cycle plan data, and lane steering function data reported by the radar video all-in-one device, correlate and fuse the above data, combine the equipment, lane, and signal timing data, group them by equipment and lane, and sort them by time field to form a table. For the corresponding relationship between equipment and lanes, see Figure 2 For the relationship between phase sequence and lane, see Figure 4 ( Figure 4 Phase A through lane is Figure 2 Device dev001 captured lanes 02 and 03. Device dev001 detected two lanes, 02 and 03. The results of device data fusion are shown in Tables 1 and 2:

[0071] Table 1

[0072]

[0073] Table 2

[0074]

[0075]

[0076] (2) Data integrity check

[0077] Perform an integrity check on the fused data. The device reports at a rate of 1 item per second. The green light duration for this lane in cycle number 503 is 31 seconds. The theoretical number of items reported by the device during this period is 31. Based on experience, the integrity threshold is set at 10% of the theoretical number of items reported, or 3 items. According to Table 1, lane 02 actually reported 29 items, with an error of 2 items. The data integrity check for lane 02 passed. According to Table 2, lane 03 actually reported 28 items, with an error of 3 items. The data integrity check for lane 03 passed.

[0078] (3) Data abnormality test

[0079] After the data integrity test is passed, the data abnormality test is carried out. The test process is as follows:

[0080] ① Headway Threshold Verification: Based on experience, the headway for saturated through lanes at intersection entrances is set at 1.5 seconds per vehicle. The time difference between any adjacent detected targets is calculated based on Tables 1 and 2. If the time difference is less than the threshold of 1.5 seconds per vehicle, the data is considered abnormal and a field is added to Tables 1 and 2 to mark it. The results of the headway threshold verification for lanes 02 and 03 are shown in Tables 3 and 4:

[0081] Table 3

[0082]

[0083]

[0084] Table 4

[0085]

[0086] ② Traffic flow parameter consistency test: Based on ①, the normal data in Tables 3 and 4 were used to calculate the traffic volume and maximum queue length, and the traffic volume and maximum queue length in the same lane were verified. Considering the saturated headway, the average vehicle length was set at 7 meters per vehicle based on empirical values. At the same time, considering the inherent error factors of the radar video integrated device in detecting queue length, the consistency difference threshold between queue length and traffic volume was set at 3 vehicles. The maximum queue length in lane 02 was 75 meters. The number of vehicles passing through during this period was 11, with an error of approximately 1 vehicle, which was less than the threshold of 3 vehicles. The traffic flow parameter consistency test for lane 02 passed. The maximum queue length in lane 03 was 40 meters. The number of vehicles passing through during this period was 9, with an error of approximately 5 vehicles, which was less than the threshold of 3 vehicles. The traffic flow parameter consistency test for lane 03 passed.

[0087] ③ Traffic flow distribution pattern verification: Lanes 02 and 03 are through lanes with the same entrance and function as the lanes, so theoretically, their traffic flow distribution patterns should be consistent. Based on experience, the queue length consistency difference threshold is set at 3 vehicles, and the flow consistency difference is set at 3 vehicles. During this cycle, 11 vehicles passed through lane 02; 9 vehicles passed through lane 03; the vehicle distribution patterns are consistent. The maximum queue length in lane 02, converted to the number of vehicles, is 10; while the maximum queue length in lane 03, converted to the number of vehicles, is 4; the queue length distribution patterns are inconsistent. Based on the above analysis, the traffic data from this device is abnormal.

[0088] (4) Data anomaly marking

[0089] The radar and video integrated device's traffic data anomaly detection method, presented herein, examines raw data. However, during signal optimization on the signal control platform, the input data is a post-settlement indicator of the raw data, measured in cycles. Common indicators include cycle flow and maximum queue length. Relying on the data anomaly detection results of the present invention to mark settlement indicators, signal optimization can prevent abnormal data from interfering with its effectiveness.

[0090] Example 2

[0091] Reference Figure 5 This embodiment relates to a system for implementing a method for detecting abnormal values ​​in traffic data of a radar video integrated device according to embodiment 1, comprising:

[0092] The data acquisition and fusion module is used to obtain traffic data, signal cycle plan data, and lane steering function data reported by the radar video integrated device, and correlate and fuse the above data;

[0093] The data integrity check module is used to judge the data omission and duplication of the fused original data. If the number of omissions and duplications of the original data does not exceed the threshold, it is marked as normal data; otherwise, it is marked as abnormal data.

[0094] The data anomaly inspection module is used to perform anomaly inspection on the basis of the completeness of the data reported by the equipment. Specifically, it includes the headway threshold inspection, traffic flow parameter consistency inspection, and traffic flow distribution pattern inspection. The data that meets the inspection standards during the inspection period is recorded as normal data, and the data that does not meet the inspection standards is recorded as abnormal data.

[0095] The data anomaly marking module is used to mark the periodic indicators of the signal control platform settlement based on the analysis results of the abnormalities of the original data reported by the equipment.

[0096] Example 3

[0097] This embodiment relates to a computer-readable storage medium having a program stored thereon. When the program is executed by a processor, the method for detecting abnormal values ​​in traffic data of a radar video integrated machine of embodiment 1 is implemented.

[0098] Example 4

[0099] This embodiment relates to a computing device, including a memory and a processor, wherein the memory stores executable code, and when the processor executes the executable code, a radar video all-in-one machine traffic data anomaly detection method of Example 1 is implemented.

Claims

1. A method for detecting abnormal values ​​in traffic data of a radar video integrated device, the steps of which are as follows: Step (1): Acquire and fuse data; obtain traffic data, signal cycle plan data, and lane steering function data reported by the radar video all-in-one device, and correlate and fuse the above data; Step (2): Data integrity check; The fused raw data is judged for missed and repeated reporting, and the raw data with missed or repeated reporting numbers not exceeding the threshold is marked as normal data, otherwise it is marked as abnormal data; Step (3): Data abnormality test; perform abnormality test on the basis of complete data reported by the equipment, specifically including headway threshold test, traffic flow parameter consistency test, and traffic flow distribution law test. The data that meets the test standards during the test period is recorded as normal data, and the data that does not meet the test standards is recorded as abnormal data. The headway threshold test is to determine whether the headway of vehicles passing through the same lane at adjacent moments is reasonable. The average headway refers to the average time difference between adjacent vehicles passing a certain section, and the unit is s / veh. The calculation formula is: t h =3600 / q (2) In the above formula: t h —headway (s / veh); q—flow rate, the number of vehicles passing through a section during the observation period (veh / h); When the flow rate q reaches the saturation flow rate, that is, the maximum number of vehicles passing through a section per unit time, the average headway is minimized, which is the saturated headway. The headway threshold should be less than the saturated headway. When the headway between adjacent vehicles is greater than the headway threshold, the headway is considered normal; otherwise, it is considered abnormal. The judgment formula is: In the above formula: —The moment when the kth vehicle passes by lane j monitored by radar and video integrated device i; —The moment when the k+1th vehicle passes by lane j monitored by radar and video integrated device i; t hmin —Threshold of headway; The traffic flow parameter consistency check verifies the consistency of different parameters reported by devices monitoring the same lane. The radar and video integrated device detects vehicles through video and the queue length through radar. The queue length reported and the number of vehicles passing through in the same cycle should be consistent. That is, the difference between the queue length converted into the number of vehicles and the actual number of vehicles in that cycle should not exceed the threshold. In this case, the traffic flow parameters are consistent. Otherwise, the traffic flow parameters are inconsistent and an anomaly is determined. The judgment formula is: In the above formula: —The maximum queue length (meters) of lane j reported by radar video integrated device i during the cycle; — Radar and video integrated device i reports data and counts the number of vehicles in lane j during the cycle; —Average vehicle length (m); N2—the consistency difference threshold between queue length and traffic; The traffic flow distribution pattern test is to check whether the lanes with the same turning direction at the same entrance of the road intersection have consistent traffic flow distribution patterns. The difference in traffic parameters between lanes with the same turning direction at the same entrance reported by the device should not exceed the threshold. That is, if the difference in the number of vehicles passing through and the maximum queue length value between multiple lanes with the same turning direction at the same entrance within a period is within the threshold, it is judged as normal data; otherwise, it is judged as abnormal data. The judgment formula is: In the above formula: —The maximum queue length (meters) of lane j reported by radar video integrated device i during the cycle; —The maximum queue length (in meters) of lane j+1 in this cycle, reported by radar video integrated device i, which has the same entrance and turns as lane j; —Average vehicle length (m); N3—the threshold value for the consistency difference in queue length between the same turning lanes at the same entrance (vehicles); In the above formula: —The number of vehicles in lane j reported by radar and video integrated device i during the cycle; — Radar and video integrated device i reports the number of vehicles in lane j+1 with the same entrance and turn as lane j in this cycle; N4—Flow consistency difference threshold between the same turning lanes of the same entrance (vehicles); Step (4): Marking of data anomalies; Based on the anomalies of the original data reported by the equipment, the periodic indicators of the signal control platform settlement are marked based on the analysis results.

2. The method for detecting abnormal values ​​in flow data of a radar video integrated machine according to claim 1, characterized in that: Step (2) of the data integrity check includes two aspects: data omission and data re-reporting. The judgment basis is to compare the theoretical value with the actual value. When the difference is within the threshold, it is marked as normal data, otherwise it is marked as abnormal data. The specific formula is: |m i ·t ig -n i |≤N1 (1) In the above formula: μ i — Frequency of data reporting of radar video integrated device i (items / s); t ig —The maximum green light duration (s) of the phase where the radar video integrated device i monitors the lane; n i —The number of data items actually reported by radar video integrated device i; N1—Threshold for missed and repeated reporting data.

3. The method for detecting abnormal values ​​in flow data of a radar video integrated machine according to claim 1, characterized in that: The data anomaly marking in step (4) is based on the abnormality inspection of the original data reported by the equipment, and marks the periodic flow and periodic maximum queue length indicators settled by the signal control platform.

4. A system for implementing the method for detecting abnormal values ​​in flow data of a radar video integrated device according to claim 1, characterized in that: include: The data acquisition and fusion module is used to obtain traffic data, signal cycle plan data, and lane steering function data reported by the radar video integrated device, and correlate and fuse the above data; The data integrity check module is used to judge the data omission and duplication of the fused original data. If the number of omissions and duplications of the original data does not exceed the threshold, it is marked as normal data; otherwise, it is marked as abnormal data. The data anomaly inspection module is used to perform anomaly inspection based on the completeness of the data reported by the equipment. Specifically, it includes headway threshold inspection, traffic flow parameter consistency inspection, and traffic flow distribution pattern inspection. Data that meets the inspection standards during the inspection period is recorded as normal data, and data that does not meet the inspection standards is recorded as abnormal data. The headway threshold inspection is to determine whether the headway of vehicles passing through the same lane at adjacent times is reasonable. The average headway refers to the average time difference between adjacent vehicles passing a certain section, and the unit is s / veh. The calculation formula is: t h =3600 / q (2) In the above formula: t h —headway (s / veh); q—flow rate, the number of vehicles passing through a section during the observation period (veh / h); When the flow rate q reaches the saturation flow rate, that is, the maximum number of vehicles passing through a section per unit time, the average headway is minimized, which is the saturated headway. The headway threshold should be less than the saturated headway. When the headway between adjacent vehicles is greater than the headway threshold, the headway is considered normal; otherwise, it is considered abnormal. The judgment formula is: In the above formula: —The moment when the kth vehicle passes by lane j monitored by radar and video integrated device i; —The moment when the radar video integrated device i monitors lane j and the k+1th vehicle passes by; t hmin —Threshold of headway; The traffic flow parameter consistency check verifies the consistency of different parameters reported by devices monitoring the same lane. The radar and video integrated device detects vehicles through video and the queue length through radar. The queue length reported and the number of vehicles passing through in the same cycle should be consistent. That is, the difference between the queue length converted into the number of vehicles and the actual number of vehicles in that cycle should not exceed the threshold. In this case, the traffic flow parameters are consistent. Otherwise, the traffic flow parameters are inconsistent and an anomaly is determined. The judgment formula is: In the above formula: —The maximum queue length (meters) of lane j in the cycle reported by radar video integrated device i; — Radar and video integrated device i reports data and counts the number of vehicles in lane j during the cycle; —Average vehicle length (m); N2—the consistency difference threshold between queue length and traffic; The traffic flow distribution pattern test is to check whether the lanes with the same turning direction at the same entrance of the road intersection have consistent traffic flow distribution patterns. The difference in traffic parameters between lanes with the same turning direction at the same entrance reported by the device should not exceed the threshold. That is, if the difference in the number of vehicles passing through and the maximum queue length value between multiple lanes with the same turning direction at the same entrance within a period is within the threshold, it is judged as normal data; otherwise, it is judged as abnormal data. The judgment formula is: In the above formula: —The maximum queue length (meters) of lane j in the cycle reported by radar video integrated device i; —The maximum queue length (in meters) of lane j+1 in this cycle, reported by radar video integrated device i, which has the same entrance and turns as lane j; —Average vehicle length (m); N3—the threshold value for the consistency difference in queue length between the same turning lanes at the same entrance (vehicles); In the above formula: —The number of vehicles in lane j reported by radar and video integrated device i during the cycle; — Radar and video integrated device i reports the number of vehicles in lane j+1 with the same entrance and turn as lane j in this cycle; N4—Flow consistency difference threshold between the same turning lanes of the same entrance (vehicles); The data anomaly marking module is used to mark the periodic indicators of the signal control platform settlement based on the analysis results of the abnormalities of the original data reported by the equipment.

5. A computer-readable storage medium, characterized in that A program is stored thereon, and when the program is executed by a processor, a method for detecting abnormal values ​​in flow data of a radar video integrated machine according to any one of claims 1 to 3 is implemented.

6. A computing device comprising a memory and a processor, wherein: The memory stores executable code, and when the processor executes the executable code, it implements a radar video integrated machine traffic data abnormal value detection method according to any one of claims 1 to 3.

Images