Noise removal device, noise removal method, and program

Abstract

A noise removal device, a noise removal method, and a computer-readable medium capable of reducing the effect of noise contained in waterfall data are provided. The noise removal device includes a determination unit (241) that determines end coordinates of each of a plurality of trajectories extracted from waterfall data, the waterfall data being acquired by measuring signals from a plurality of sensing units arranged along a road, an extension unit (242) that uses the end coordinates to extend each trajectory so that the duration of each trajectory becomes a reference duration, a feature calculation unit (243) that calculates features of each extended trajectory, a detection unit (244) that detects an outlier trajectory from the plurality of trajectories based on the features, and a removal unit (245) that removes the outlier trajectory from the plurality of trajectories.

Description

[Technical Field]

[0001] The present disclosure relates to a noise removal apparatus, a noise removal method, and a computer-readable medium. [Background technology]

[0002] Optical fibers are sometimes laid along roads (e.g., highways). The optical fibers are equipped with multiple sensing units along the road. Distributed Acoustic Sensing (DAS) attached to the optical fibers can detect vibrations at the locations where each sensing unit is installed.

[0003] Distributed acoustic sensors acquire data called waterfall data, which includes information about the time and location at which vibrations were detected. Based on the waterfall data, the speed of a vehicle traveling on a road or the like can be calculated (see Patent Document 1). [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Publication No. 2021-121917 Summary of the Invention [Problem to be solved by the invention]

[0005] If the waterfall data contains noise, the accuracy of calculating the vehicle's average speed decreases.

[0006] In view of the above circumstances, an object of the present disclosure is to provide a noise removal device, a noise removal method, and a computer-readable medium that reduce the effect of noise contained in waterfall data. [Means for solving the problem]

[0007] The present disclosure provides: a determining means for determining end coordinates of each of a plurality of trajectories extracted from waterfall data, the waterfall data being obtained by measuring signals from each of a plurality of sensing units arranged along a road; an extension means for extending each locus using the end coordinates so that the duration of each locus becomes a reference duration; a feature calculation means for calculating the features of each extended trajectory; a detection means for detecting an outlier trajectory from the plurality of trajectories based on the characteristics; a removal means for removing the outlier trajectory from the plurality of trajectories; The present invention provides a noise removal device comprising:

[0008] The present disclosure provides: determining end coordinates of each of a plurality of trajectories extracted from waterfall data, the waterfall data being obtained by measuring signals from each of a plurality of sensing units arranged along a road; Extending each trajectory using the end coordinates so that the duration of each trajectory becomes a reference duration; calculating characteristics of each extended trajectory; a detection unit that detects an outlier trajectory from the plurality of trajectories based on the features; removing the outlier trajectories from the plurality of trajectories; The present invention provides a noise removal method including:

[0009] The present disclosure provides: determining end coordinates of each of a plurality of trajectories extracted from waterfall data, the waterfall data being obtained by measuring signals from each of a plurality of sensing units arranged along a road; Extending each trajectory using the end coordinates so that the duration of each trajectory becomes a reference duration; calculating characteristics of each extended trajectory; detecting an outlier trajectory from the plurality of trajectories based on the features; removing the outlier trajectories from the plurality of trajectories; A non-transitory computer-readable medium is provided that stores a program that causes a computer to execute processes including the steps of: [Effects of the Invention]

[0010] The noise removal device, noise removal method, and computer-readable medium according to the present disclosure reduce the effect of noise contained in waterfall data. [Brief explanation of the drawings]

[0011] [Figure 1] 1 is a schematic diagram showing a traffic monitoring system according to a first embodiment. [Figure 2] 1 is a flowchart showing the flow of a traffic monitoring method according to a related art. [Figure 3] Schematic diagram showing a patch. [Figure 4] Schematic diagram showing noise contained in a patch. [Figure 5] 1 is a table showing average speeds calculated using related techniques. [Figure 6] 10A and 10B are schematic diagrams illustrating trajectories included in a patch. [Figure 7] FIG. 1 is a block diagram showing a traffic monitoring device according to a first embodiment. [Figure 8] FIG. 10 is a diagram illustrating a method for extending a trajectory. [Figure 9] FIG. 10 is a diagram for explaining the reason for extending the trajectory. [Figure 10] FIG. 10 is a diagram for explaining the reason for extending the trajectory. [Figure 11] FIG. 10 is a diagram for explaining the reason for extending the trajectory. [Figure 12] Scatter plot with points corresponding to the extended trajectories. [Figure 13] Scatter plot of road traffic conditions as they change. [Figure 14] Scatter plot of road traffic conditions as they change. [Figure 15] 1 is a flowchart showing the flow of a noise removal method according to the first embodiment. [Figure 16] 3 is a flowchart showing an example of the flow of a noise removal method according to the first embodiment. [Figure 17] FIG. 2 is a diagram for explaining the effect of the first embodiment. [Figure 18] FIG. 2 is a diagram for explaining the effect of the first embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0012] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings. In each drawing, the same or corresponding elements are designated by the same reference numerals, and for clarity of explanation, duplicate explanations will be omitted as necessary.

[0013] (Embodiment 1) 1 is a schematic diagram of a distributed acoustic sensor (DAS) system 1000 along a road 10. The DAS system 1000 includes a traffic monitoring device 200 in communication with the DAS 100. The traffic monitoring device 200 is also referred to as a noise filter. The DAS system 1000 further includes an optical fiber 300 connected to the DAS 100. The optical fiber 300 runs along the road 10.

[0014] The road 10 may be a highway. The road 10 may include multiple lanes. Multiple vehicles 20 travel on the road 10.

[0015] The optical fiber 300 includes multiple sensing elements along the road 10. The sensing elements may be located at multiple equidistant points. When a vehicle 20 passes over the road 10, the vehicle 20 generates vibrations. These vibrations affect the propagation of light along the optical fiber 300. The DAS 100 is connected to the optical fiber 300, sends optical signals into the optical fiber 300, and detects the light returned from the optical fiber 300. The resulting data is called waterfall data. The waterfall data is a time-distance graph. Vehicle trajectories are extracted from the waterfall data, for example, by the TrafficNET algorithm. The waterfall data provides parameters such as the number of vehicles 20 on the road 10, the average vehicle speed, and lane occupancy.

[0016] Next, problems of the related art will be described with reference to Figures 2 to 6. Figure 2 is a flowchart showing the flow of a traffic monitoring method according to the related art. In the related art, first, waterfall data is acquired from the DAS 100 (step S101). Image 11 shows the waterfall data after preprocessing. The acquired raw waterfall data is preprocessed. The vertical axis represents time, and the horizontal axis represents position along the optical fiber 300. The waterfall data includes the trajectory of the vehicle 20. The vehicle trajectory is a distinguishable line that indicates vibrations caused by the vehicle 20 crossing the road 10.

[0017] Next, the trajectory of the vehicle 20 is extracted from the waterfall data (step S102). Image 12 shows the result of extracting the trajectory of the vehicle 20 from the waterfall data shown in Image 11. In Image 12, the extracted trajectory of the vehicle 20 is highlighted. The extraction of the trajectory of the vehicle 20 is performed using, for example, a DNN (Deep Neural Network) (e.g., U-Net).

[0018] Next, the average speed (km / hr) of the vehicles 20 is calculated (step S103). The speed of each vehicle is calculated based on the slope of the corresponding trajectory. Finally, traffic monitoring is performed based on the average vehicle speed (step S104). Specifically, congestion, jams, or queues may be detected. The estimated average vehicle speed, which represents the traffic flow, is the basis for traffic flow monitoring.

[0019] Next, a method for calculating the average speed of a vehicle 20 in the related art will be described with reference to FIG. 3. FIG. 3 is a schematic diagram of waterfall data. The waterfall data can be separated into multiple patches. Reference numeral 13 represents the entire waterfall data, and reference numeral 14 represents a patch. The range of the patch 14 is indicated by a dotted line in the waterfall data 13. By separating the waterfall data 13 into multiple patches 14, traffic monitoring can be performed with high accuracy. The patches are considered depending on the required flow resolution (time, distance). All trajectories within the patch 14 can be considered for the average speed.

[0020] The vertical axis represents time, and the horizontal axis represents the position along the optical fiber 300, i.e., the distance traveled by the vehicle 20. The waterfall data 13 includes multiple trajectories 15. The patch 14 includes a trajectory 15n. Δdn represents the length of the trajectory 15n in the distance direction, and Δtn represents the length of the trajectory 15n in the time direction. Δdn represents the distance traveled by the vehicle 20, and Δtn represents the elapsed time of the vehicle 20. The elapsed time is also referred to as the duration. The average vehicle speed of the patch 14 is calculated using the following formula:

[0021] Average speed = (sum of vehicle distance traveled) / (sum of duration)

[0022] FIG. 4 illustrates trajectories included in patch 14. Patch 14, surrounded by a dotted line, is included in the waterfall data. The waterfall data includes trajectories 15a1 and 15a2. Trajectory 15a1 is an actual trajectory resulting from the driving of vehicle 20. Trajectory 15a2 is a trajectory resulting from extracted noise. The noise trajectory may result from the presence of a bridge, tunnel, or other structure on the highway / road being monitored.

[0023] Figure 5 shows a table containing the results of calculating the average vehicle speed from the patch 14. Figure 5 includes the measurement date and time 21, the average vehicle speed 22 (unit: km / hr) calculated from the waterfall data, and the average vehicle speed 23 (unit: km / hr) that is the reference / ground truth data calculated from the traffic counter. The traffic counter may be a loop coil type.

[0024] When the measurement date and time 21 is T4 of D4, the average vehicle speed 22 is 150.6 km / hr and the average vehicle speed 23 is 95 km / hr. The average vehicle speed 22 and the average vehicle speed 23 are different. Noise affects the speed.

[0025] The smaller the slope of the trajectory included in patch 14, the larger the calculated average vehicle speed. Referring to FIG. 4, it is believed that the noise trajectory 15a2 with a small slope included in patch 14 causes the calculated average vehicle speed 22 to be higher than the actual speed. The noisy trajectory contributes to the calculation of the average vehicle speed, resulting in an erroneous speed. Such an erroneous speed may lead to insufficient traffic flow monitoring and reduced reliability.

[0026] Fig. 6 shows examples of trajectories included in patch 14. Fig. 6 includes trajectories 15b1, 15b2, 15b3, 15b4, and 15b5. Trajectory 15b1 is an actual trajectory resulting from vibration of vehicle 20. Trajectory 15b1 is sufficiently long, so it can be determined to be an actual trajectory.

[0027] The trajectory 15b2 is contained in the upper left side of the patch 14, and the trajectory 15b3 is contained in the lower right side of the patch 14. Due to the short lengths of the trajectories 15b2 and 15b3, the trajectories 15b2 and 15b3 may be noise or may have been cut off from the actual trajectories.

[0028] The trajectories 15b4 and 15b5 are structural noise, and because their lengths are sufficiently short, they can be determined to be noise trajectories. Such noise can be generated by structures such as bridges and tunnels along the highway. The trajectory 15b4 extends vertically, which may result in the average vehicle speed being calculated as low. The trajectory 15b5 extends horizontally, which may result in the average vehicle speed being calculated as high.

[0029] According to the related art, there is a problem that the accuracy of calculating the average vehicle speed is reduced due to the inclusion of noise in patch 14. Since short trajectories in patch 14 (e.g., trajectories 15b2 and 15b3) may be part of the actual trajectory, it is difficult to remove noise depending on the trajectory length. It is difficult to distinguish between a noise trajectory and an actual trajectory based on the trajectory length alone.

[0030] Next, the configuration of the traffic monitoring device 200 will be described with reference to Fig. 7. The traffic monitoring device 200 includes an acquisition unit 210, an extraction unit 220, a separation unit 230, a noise removal unit 240, and a traffic monitoring unit 250. The traffic monitoring device also includes a processor and a memory. Each function of the traffic monitoring device 200 can be realized by loading a program (not shown) into a memory such as RAM and executing it with the processor.

[0031] The acquisition unit 210 acquires waterfall data from the DAS 100. The extraction unit 220 extracts a plurality of trajectories from the waterfall data. The separation unit 230 separates the waterfall data into a plurality of patches.

[0032] The noise removal unit 240 removes noise from the trajectory included in each patch. The noise removal unit 240 includes a determination unit 241, an extension unit 242, a feature calculation unit 243, a detection unit 244, and a removal unit 245.

[0033] The determination unit 241 determines the end coordinates of each of the multiple trajectories extracted from the waterfall data. As described above, the waterfall data is acquired by measuring signals from multiple sensing units arranged along the road 10. The determination unit 241 outputs the determined end coordinates to the extension unit 242.

[0034] The extension unit 242 uses the end coordinates to extend each of the multiple trajectories so that the duration of each trajectory becomes the reference duration. The reference duration may be a fixed time.

[0035] Next, the extension process performed by the extension unit 242 will be specifically described with reference to Fig. 8. The left side of Fig. 8 shows the patch 14 before the extension process is performed. The patch 14 includes multiple trajectories. The patch 14 includes a short trajectory (for example, trajectory 15c1). The trajectory 15c1 includes end points E1 and E2.

[0036] The right side of FIG. 8 shows the patch 14 after the extension process. The extension unit 242 extends each trajectory so that the duration of each trajectory becomes the reference duration T. All trajectories within the patch 14 can be extended. The reference duration T matches the time width of the patch 14. Note that the reference duration and the time width of the patch 14 may differ. For ease of viewing, the portions of each extended trajectory that are included inside the patch 14 are shown with solid lines. The portions of each extended trajectory that are included outside the patch 14 are shown with dotted lines. Trajectory 15c2 shows a trajectory that is an extension of trajectory 15c1.

[0037] As mentioned above, it is difficult to classify each trajectory into an actual trajectory and noise trajectory based on its length. This is because the actual trajectory and the noise trajectory may have the same length. The traffic monitoring device 200 extends the trajectory up to a reference duration T to fit all trajectories into a common frame. This allows the length of each trajectory having the reference duration T to be determined. The length may be the length of the line or the length in the distance direction. The length of each trajectory differs between the actual trajectory and the noise trajectory. This allows the removal unit 245, which will be described later, to remove outlier trajectories (noise trajectories).

[0038] Next, with reference to FIGS. 9 to 11, the reason why the extension unit 242 performs trajectory extension will be described in detail. The patch 14 shown in FIG. 9 includes trajectories 15d1, 15d2, 15d3, 15d4, and 15d5. Trajectories 15d1, 15d2, 15d3, and 15d4 are noise trajectories. Trajectory 15d5 is a true trajectory resulting from vehicle vibration. Because the noise trajectories are each at a different reference time, extension is performed to align all trajectories to a common reference. Since trajectory 15d5 is longer than the other noise trajectories 15d1, 15d2, 15d3, and 15d4, it can be said that longer trajectories are more likely to be true vehicle trajectories. Therefore, true vehicle trajectories have a higher extension ratio / weight.

[0039] FIG. 10 shows the length of each trajectory included in patch 14 in the time direction. Δt1 indicates the duration of trajectory 15d2, Δt2 indicates the duration of trajectory 15d3, Δt3 indicates the duration of trajectory 15d4, and Δt4 indicates the duration of trajectory 15d5. The durations are the lengths of the trajectories in the time direction. Δt1, Δt2, Δt3, and Δt4 are different from one another. It is not possible to identify noise trajectories (outlier trajectories) based on the lengths of trajectories 15d2, 15d3, 15d4, and 15d5. Therefore, the traffic monitoring device 200 extends each trajectory so that its duration becomes the reference duration T.

[0040] Figure 11 shows the entire waterfall data set 13, including patch 14. The horizontal axis represents distance (0-15 km), and the vertical axis represents time. By comparing the extended trajectory with the waterfall data set 13, the reliability of the trajectory can be calculated. For example, an extended trajectory of trajectory 15d5 may overlap with a trajectory in the waterfall data set 13. In this case, trajectory 15d5 is likely to originate from actual vehicle vibration. Trajectory 15d5 is a discontinuous vehicle trajectory and is part of a larger trajectory. Therefore, the trajectory needs to be extended. Small patches lose context, and larger patches are required for outlier detection.

[0041] Returning to FIG. 7, the explanation continues. The feature calculation unit 243 calculates the feature of each extended trajectory. The feature is, for example, the length of each extended trajectory. The length of each extended trajectory may be the length of a line for a certain time (reference duration). The speed of a vehicle within a predetermined distance range and time range is considered to be within a certain range. Therefore, noise can be removed by calculating a parameter related to the vehicle speed from each trajectory. The length of an extended trajectory in the distance direction is the product of the speed and the reference duration T, and is therefore related to the speed. Furthermore, as will be explained below, the length of an extended trajectory is also related to the speed.

[0042] Next, referring to FIG. 12, it will be explained that the length of the extended trajectory can be used as a feature. FIG. 12 is a scatter plot in which points corresponding to the extended trajectories are plotted, with the vertical axis representing the length of the extended trajectory and the horizontal axis representing the velocity corresponding to the trajectory. The points correspond to all trajectories within the patch. The velocity is calculated based on the slope of the trajectory. There is a correlation between the velocity and the length of the extended trajectory, and the longer the trajectory, the greater the velocity. Therefore, the length of the extended trajectory can be used as a feature.

[0043] The scatter plot includes regions 31, 32, and 33. Region 31 includes outlier trajectories on the slow side. Region 31 includes, for example, extended trajectory 15e1 shown on the left. The length of extended trajectory 15e1 is smaller than the length of extended trajectory 15e2, which will be described later.

[0044] Region 32 includes points corresponding to the actual vehicle trajectory. In region 32, the density of points corresponding to the actual vehicle trajectory is high. Region 32 includes extended trajectories of average length. Region 32 includes, for example, extended trajectory 15e2 shown on the left. The length of extended trajectory 15e2 is greater than the length of extended trajectory 15e1 and less than the length of extended trajectory 15e3, which will be described later.

[0045] Region 33 includes outlier trajectories on the high-speed side. Region 33 includes, for example, extended trajectory 15e3 shown on the left. The length of extended trajectory 15e3 is greater than the length of extended trajectory 15e2.

[0046] Figure 12 shows the distribution of the length of the extended trajectories in normal traffic. As traffic changes, the distribution of the length of the extended trajectories changes. Figure 13 shows the distribution of the length of the extended trajectories when the vehicle (speeding vehicle) is moving at a high speed, where the region 32 shifts to the upper right. Figure 14 shows the distribution of the length of the extended trajectories when traffic congestion occurs, where the region 32 shifts to the lower left.

[0047] 7, the description will be continued. The detection unit 244 detects outlier trajectories from the plurality of trajectories based on the features. The removal unit 245 removes the outlier trajectories from the plurality of trajectories.

[0048] The traffic monitoring unit 250 calculates the average vehicle speed for each patch. The traffic monitoring unit 250 calculates the speed of vehicles traveling on the road based on the gradient of each trajectory, and calculates the average speed. The traffic monitoring unit 250 may also calculate the average vehicle speed by taking into account the reliability of each trajectory. As explained using FIG. 11, the reliability of each trajectory can be calculated by comparing the extended trajectory with the waterfall data. Specifically, the traffic monitoring unit 250 measures the length of the overlapping portion between each extended trajectory and any trajectory extracted from the waterfall data. Then, the traffic monitoring unit 250 derives a weighting coefficient according to the length (length of the trajectory) and calculates the average speed based on the weighting coefficient. The weighting coefficient is also called the extension rate. The extension rate is calculated using the following formula:

[0049] Elongation ratio = (trace length) / (extended trajectory length)

[0050] The stretch ratio determines the weight of each trajectory, with longer consecutive trajectories having a higher weight in calculating the average velocity.

[0051] Next, the traffic monitoring method according to the first embodiment will be described with reference to Fig. 15. Comparing Fig. 15 with Fig. 2 showing the traffic monitoring method according to the related art, step S200 of removing noise has been added. Steps S101 to S104 are the same as Fig. 2, and therefore their explanation will be omitted. In step S200, a process is performed to remove noise loci that indicate noise from the multiple loci extracted in step S102.

[0052] Next, an example of the traffic monitoring method according to the first embodiment will be described with reference to Fig. 16. First, data D1 including a plurality of trajectories extracted from waterfall data is input.

[0053] Next, the extension unit 242 of the traffic monitoring device 200 extends each trajectory so that the duration of each trajectory becomes the reference duration (step S201). Note that before step S201, the determination unit 241 of the traffic monitoring device 200 may determine the end coordinates of each trajectory. Next, the traffic monitoring device 200 calculates the speed corresponding to each trajectory, calculates the length of each trajectory, and calculates the extension rate of each trajectory (step S202).

[0054] Next, the traffic monitoring device 200 checks whether the maximum value of the calculated speed exceeds a threshold or the minimum value of the calculated speed is less than a threshold (step S203). If the result is false (No in step S203), the process proceeds to step S104 to monitor the traffic flow. If the result is true (Yes in step S203), the process proceeds to step S204 to perform noise removal processing. The detection unit 244 of the traffic monitoring device 200 detects outlier trajectories from the multiple trajectories based on the trajectory lengths calculated in step S202. The removal unit 245 of the traffic monitoring device 200 removes the outlier trajectories from the multiple trajectories. In this case, traffic monitoring is performed using trajectory data D2 from which the outlier trajectories have been removed (step S104). The traffic monitoring unit 250 of the traffic monitoring device 200 may calculate an average speed based on the data D2 and monitor the traffic flow.

[0055] Next, the effects of the traffic monitoring device 200 will be described with reference to FIGS. 17 and 18. FIG. 18 includes waterfall data 13a1 acquired by the DAS 100. The waterfall data 13a1 includes sudden changes in traffic flow (e.g., trajectory 15f1). For example, a portion of the trajectory 15f1 with a steep slope corresponds to a traffic congestion. However, the waterfall data 13a1 also includes outlier trajectories (e.g., trajectory 15f2), and if the average vehicle speed is measured for each patch, it may not be possible to detect the traffic congestion. The traffic monitoring device 200 executes algorithm A1 to remove noise from the waterfall data 13a1.

[0056] The waterfall data 13a2 shows data after noise has been removed. By executing the traffic monitoring algorithm A2 based on the waterfall data 13a2, a sudden change in traffic flow (e.g., traffic congestion) can be detected.

[0057] Fig. 18 shows the calculation results of the average vehicle speed when the related technology is used and the calculation results of the average vehicle speed when the embodiment 1 is used. The upper part of Fig. 18 includes the processing flow when the related technology is used, and the lower part includes the processing flow when the embodiment 1 is used.

[0058] Patch 14a1 shows a patch before noise removal. Patch 14a1 contains multiple noise trajectories. In the related art, the average vehicle speed V (e.g., 150.6 km / hr) is calculated from the entire trajectory, including the noise trajectories. In such a case, applying algorithm A2, which detects sudden changes in traffic flow, can result in a false alarm. Table T1 shows the results of calculating the average speed using the related art.

[0059] Meanwhile, the traffic monitoring device 200 applies the noise removal algorithm A2 to patch 14a1. Patch 14a2 shows the patch after noise has been removed from patch 14a1. The traffic monitoring device 200 calculates the average speed V (e.g., 95 km / hr) from the trajectory that does not include the noise trajectory. In this case, applying algorithm A2, which detects sudden changes in traffic flow, will prevent false alarms from being output. Table T2 shows the results of the average speed calculated by the traffic monitoring device 200.

[0060] According to the first embodiment, the accuracy of estimating the average speed is improved. In addition, events such as accidents and traffic jams can be detected with high efficiency, and false alarms can be reduced.

[0061] The program includes instructions (or software code) that, when loaded into a computer, cause the computer to perform one or more functions described in the embodiments. The program may be stored in a non-transitory computer-readable medium or a tangible storage medium. By way of example and not limitation, computer-readable media or tangible storage media include random-access memory (RAM), read-only memory (ROM), flash memory, solid-state drive (SSD) or other memory technologies, CD-ROM, digital versatile disc (DVD), Blu-ray® disc or other optical disk storage, magnetic cassette, magnetic tape, magnetic disk storage or other magnetic storage devices. The program may also be transmitted on a transitory computer-readable medium or communication medium. By way of example and not limitation, transitory computer-readable media or communication media include electrical, optical, acoustic, or other forms of propagated signals.

[0062] The above describes the embodiments of the present disclosure in detail, but the present disclosure is not limited to the above-described embodiments, and changes and modifications to the above-described embodiments that do not deviate from the spirit of the present disclosure are also included in the present disclosure. [Explanation of symbols]

[0063] 1000: Traffic monitoring system 100:DAS 200: Traffic monitoring device 210: Acquisition Department 220: Extraction part 230: Separation part 240: Noise removal section 241: Decision Section 242: Extension part 243: Feature calculation unit 244:Detection unit 245:Removal section 250: Traffic monitoring department 300: Optical fiber 10: Road 20: Vehicle 11, 12: Images 13, 13a1, 13a2: Waterfall Data 14: Patch 15, 15n, 15a1, 15a2, 15b1, 15b2, 15b3, 15b4, 15b5, 15c1, 15c2, 15d1, 15d2, 15d3, 15d4, 15d5, 15e1, 15e2, 15e3, 15f1, 15f2: Locus 21: Measurement date and time 22:Average vehicle speed 23:Average vehicle speed 31, 32, 33: area A1, A2: Algorithm E1, E2: End point coordinates T: Reference duration T1, T2: table

Claims

1. a determining means for determining end coordinates of each of a plurality of trajectories extracted from waterfall data, the waterfall data being obtained by measuring signals from each of a plurality of sensing units arranged along a road; an extension means for extending each locus using the end coordinates so that the duration of each locus becomes a reference duration; a feature calculation means for calculating the features of each extended trajectory; a detection means for detecting an outlier trajectory from the plurality of trajectories based on the characteristics; a removal means for removing the outlier trajectory from the plurality of trajectories; a traffic monitoring means for calculating the speed of a vehicle traveling on the road based on the slope of each trajectory and for calculating the average speed of vehicles traveling on the road; Equipped with the traffic monitoring means measures the length of a portion of each extended trajectory that overlaps with any of the trajectories extracted from the waterfall data, derives a weighting coefficient according to the length of the portion, and calculates the average speed based on the weighting coefficient; Noise removal device.

2. the noise removal device includes a separation means for separating the waterfall data into a plurality of patches; the determining means determines the end coordinates of each locus included in each patch; the traffic monitoring means calculates the average speed for each patch; The noise removal device according to claim 1 .

3. The feature is the length of each trajectory after extension. The noise removal device according to claim 1 or 2.

4. A computer comprising: determining end coordinates of each of a plurality of trajectories extracted from waterfall data, the waterfall data being obtained by measuring signals from each of a plurality of sensing units arranged along a road; Extending each trajectory using the end coordinates so that the duration of each trajectory becomes a reference duration; calculating characteristics of each extended trajectory; detecting an outlier trajectory from the plurality of trajectories based on the features; removing the outlier trajectories from the plurality of trajectories; Calculating the speed of a vehicle traveling on the road based on the slope of each trajectory, and calculating an average speed of the vehicles traveling on the road; Including, Calculating the average speed includes measuring a length of a portion of each of the extended trajectories that overlaps with any of the trajectories extracted from the waterfall data, deriving a weighting coefficient according to the length of the portion, and calculating the average speed based on the weighting coefficient. Noise removal method.

5. determining end coordinates of each of a plurality of trajectories extracted from waterfall data, the waterfall data being obtained by measuring signals from each of a plurality of sensing units arranged along a road; Extending each trajectory using the end coordinates so that the duration of each trajectory becomes a reference duration; calculating characteristics of each extended trajectory; detecting an outlier trajectory from the plurality of trajectories based on the features; removing the outlier trajectories from the plurality of trajectories; Calculating the speed of a vehicle traveling on the road based on the slope of each trajectory, and calculating an average speed of the vehicles traveling on the road; causing a computer to execute a process including Calculating the average speed includes measuring a length of a portion of each of the extended trajectories that overlaps with any of the trajectories extracted from the waterfall data, deriving a weighting coefficient according to the length of the portion, and calculating the average speed based on the weighting coefficient. program.

Images