DAS and dynamic threshold based highway low speed event monitoring method and system
By using a DAS-based dynamic threshold monitoring method, the timeliness and adaptability issues of traditional monitoring methods are solved, enabling accurate identification and timely alarm of low-speed events on highways, and reducing the false alarm and missed alarm rates.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- ANHUI HUANYU HIGHWAY CONSTR & DEV CO LTD
- Filing Date
- 2026-02-12
- Publication Date
- 2026-06-09
AI Technical Summary
Traditional methods for monitoring low-speed events on highways rely on manual reporting, which has low timeliness. Fixed threshold algorithms have poor adaptability, resulting in high false alarm and false alarm rates. They are unable to adapt to the judgment of low-speed events at different times and locations.
A dynamic threshold monitoring method based on DAS is adopted. By preprocessing the DAS signal, enhancing, separating and detecting the vehicle trajectory signal, and combining it with Hough transform, the average speed of vehicles in the section is obtained, and low-speed events are judged by using dynamic baseline threshold.
It significantly reduced the false alarm and missed alarm rates for low-speed events, improved the timeliness of event handling, and reduced economic losses.
Smart Images
Figure CN122176935A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of highway low-speed event monitoring technology, and in particular to a highway low-speed event monitoring method and system based on DAS and dynamic thresholds. Background Technology
[0002] Currently, the traditional monitoring of low-speed incidents on highways mainly relies on manual reporting. This method is often not very timely. When an incident is reported quickly, the monitoring center needs about 10 minutes to receive the report. If the incident is reported slowly, the monitoring center will need about 30 minutes to receive the report. This will greatly reduce the overall efficiency of incident handling, greatly affect the subsequent traffic conditions on the highway, and cause greater economic losses.
[0003] The traditional algorithm is the fixed threshold algorithm, which sets a fixed speed limit. If the speed is below this limit, it is judged as a low-speed event. Although this algorithm makes up for the timeliness problem of manual reporting, it has the following problems: the fixed threshold algorithm has poor adaptability to the judgment of low-speed events at different times and locations on highways, and the false alarm rate and missed alarm rate are relatively high.
[0004] For example, at the same location at different times, if a fixed threshold is set for the daytime, false alarms may occur at night when vehicle speeds are lower. If a lower fixed threshold is used, false alarms may occur. At different locations at the same time, some road sections have higher vehicle speeds than others. If a uniform fixed threshold is set, more false alarms will occur for road sections with lower vehicle speeds, and more false alarms will occur for road sections with higher vehicle speeds. If a fixed threshold is set for different road sections based on their vehicle speeds, the algorithm complexity will be greatly increased, and the problems mentioned above regarding different times at the same location will also occur. Summary of the Invention
[0005] To address the technical problems existing in the background art, this invention proposes a method and system for monitoring low-speed events on highways based on DAS and dynamic thresholds.
[0006] In a first aspect, the present invention proposes a method for monitoring low-speed events on highways based on DAS and dynamic thresholds, comprising: Acquire DAS signals for the target highway segment; The DAS signal is preprocessed to obtain a feature image including vehicle trajectory signal; Enhance vehicle trajectory signals in feature images; Vehicle trajectory signals in different directions are separated in the enhanced feature image to obtain vehicle trajectory separation feature maps in different directions; Based on the segmentation of the target highway segment, the vehicle trajectory separation feature map for different directions is divided into segments to obtain multiple segment feature maps; Edge detection and Hough transform are performed on the feature maps of each segment to obtain the coordinates of the straight line segments of the vehicle trajectory of each vehicle in each segment. The average speed of all vehicles in each segment is obtained by using the endpoint coordinates of the straight line segments of each vehicle's trajectory within each segment. Based on the average speed of all vehicles in each segment and the dynamic baseline threshold of that segment, determine whether a low-speed event exists in that segment; if so, issue an alarm.
[0007] Preferably, the preprocessing includes: common-mode noise removal, bad sector removal, and filtering.
[0008] Preferably, the feature image is enhanced with vehicle trajectory signals, specifically including: Perform absolute value processing on the feature image; The feature image after absolute value processing is normalized by interval summation.
[0009] Preferably, the vehicle trajectory signals in different directions in the enhanced feature image are separated to obtain a vehicle trajectory separation feature map, specifically including: The enhanced feature image data is transformed from the spatiotemporal domain to the frequency-wavenumber domain by Fourier transform, and the logarithmic amplitude spectrum and phase spectrum of the feature image are extracted. In the frequency domain-wavenumber domain, the logarithmic amplitude spectrum and phase spectrum of the feature image are subjected to sector filtering according to polar coordinate angles in different preset directions; The filtered logarithmic amplitude spectrum and phase spectrum are converted from the frequency domain to the wavenumber domain and back to the spatiotemporal domain by inverse Fourier transform, and the data with corresponding angles are reconstructed to obtain vehicle trajectory separation feature maps in different directions.
[0010] Preferably, the polar coordinate angles are 0-30° and 180°-210°, or 150°-180° and 330°-360°.
[0011] Preferably, edge detection and Hough transform are performed on the feature maps of each segment to obtain the coordinates of the straight line segments of the vehicle trajectory of each vehicle in each segment, specifically including: The Canny algorithm is used to separate the edges of vehicle trajectories in the feature maps of each segment, resulting in a binarized edge image for each segment. By using Hough transform, the vehicle trajectory line segments are extracted from the binarized edge image of each segment, and the coordinates of the vehicle trajectory line segments of each vehicle in each segment are obtained.
[0012] Preferably, the average speed of all vehicles in each segment is obtained based on the endpoint coordinates of the straight segments of each vehicle's trajectory within each segment, specifically including: The slope of each straight line segment is calculated based on the endpoint coordinates of the straight line segments of each vehicle trajectory within each segment. Calculate the average slope of each straight segment and use it as the average speed of all vehicles in each segment.
[0013] Preferably, the dynamic baseline threshold is the average speed of the vehicle within a first preset time period in the section; Specifically, based on the average speed of all vehicles within each segment and the dynamic baseline threshold of that segment, it is determined whether a low-speed event exists in that segment, including: If the average speed of a certain segment is consistently lower than 85% of the dynamic baseline threshold during the second preset time period, it is determined that a low-speed event exists in that segment; otherwise, it is determined that no low-speed event exists in that segment. Wherein, the sum of the first preset time period and the second preset time period is less than or equal to the total driving time of the vehicle in the section, and the first preset time period is earlier than the second preset time period in terms of time sequence, and the duration of the first preset time period is greater than the duration of the second preset time period.
[0014] Secondly, this invention also proposes a highway low-speed event monitoring system based on DAS and dynamic thresholds, comprising: The acquisition module is used to acquire the DAS signal of the target highway segment; The processing module is used to preprocess the DAS signal to obtain a feature image including vehicle trajectory signals; enhance the vehicle trajectory signals in the feature image; separate the vehicle trajectory signals in different directions in the enhanced feature image to obtain vehicle trajectory separation feature maps in different directions; divide the vehicle trajectory separation feature maps in different directions into segments according to the segment settings of the target highway segment to obtain multiple segment feature maps; perform edge detection and Hough transform on the segment feature maps to obtain the coordinates of the straight line segments of the vehicle trajectory of each vehicle in each segment; obtain the average speed of all vehicles in each segment based on the endpoint coordinates of the straight line segments of the vehicle trajectory in each segment; determine whether there is a low-speed event in the segment based on the average speed of all vehicles in each segment and the dynamic baseline threshold of the segment; if so, issue an alarm.
[0015] Preferably, it also includes an optical cable pre-laid on the target highway section and a DAS system connected to the optical cable, the DAS system being communicatively connected to the acquisition module.
[0016] The proposed method and system for monitoring low-speed events on highways based on DAS and dynamic thresholds accurately extracts vehicle trajectory signals from different directions by preprocessing DAS signals, enhancing and separating vehicle trajectory signals, and performing edge detection and Hough transform. This provides a benchmark for section speed measurement. Furthermore, the system uses dynamic baseline thresholds to adapt to low-speed events on various sections of the highway, thereby reducing the false alarm rate and the missed alarm rate of low-speed events. Attached Figure Description
[0017] Figure 1 This is a flowchart illustrating a method for monitoring low-speed events on highways based on DAS and dynamic thresholds, as proposed in one embodiment of the present invention.
[0018] Figure 2 This is a schematic diagram of the enhanced feature image in one embodiment of the present invention.
[0019] Figure 3 This is a schematic diagram of vehicle trajectory separation feature maps in different directions in one embodiment of the present invention; wherein (a) is the first preset direction and (b) is the second preset direction.
[0020] Figure 4 This is a schematic diagram of vehicle trajectory separation feature maps in different directions in one embodiment of the present invention; wherein (a) is the first preset direction and (b) is the second preset direction. Detailed Implementation
[0021] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.
[0022] Firstly, referring to Figure 1 The present invention proposes a method and system for monitoring low-speed events on highways based on DAS and dynamic thresholds, comprising: Acquire DAS signals for the target highway segment; The DAS signal is preprocessed to obtain a feature image including vehicle trajectory signal; Enhance vehicle trajectory signals in feature images; Vehicle trajectory signals in different directions are separated in the enhanced feature image to obtain vehicle trajectory separation feature maps in different directions; Based on the segmentation of the target highway segment, the vehicle trajectory separation feature map for different directions is divided into segments to obtain multiple segment feature maps; Edge detection and Hough transform are performed on the feature maps of each segment to obtain the coordinates of the straight line segments of the vehicle trajectory of each vehicle in each segment. The average speed of all vehicles in each segment is obtained by using the endpoint coordinates of the straight line segments of each vehicle's trajectory within each segment. Based on the average speed of all vehicles in each segment and the dynamic baseline threshold of that segment, determine whether a low-speed event exists in that segment; if so, issue an alarm.
[0023] This invention accurately extracts vehicle trajectory signals from different directions by preprocessing DAS signals, enhancing and separating vehicle trajectory signals, and performing edge detection and Hough transform, providing a benchmark for section speed measurement. It also adopts a dynamic baseline threshold to adapt to low-speed events on various sections of the highway, reducing the false alarm rate and false alarm rate of low-speed events.
[0024] It should be understood that the DAS signal in this embodiment is collected through optical cables pre-laid in the central green belt of the target highway section and the DAS system connected to the optical cables.
[0025] The preprocessing in this embodiment includes: common-mode noise removal, bad sector removal, and filtering.
[0026] During the common-mode noise removal process, the common noise mode of all row channels of the DAS signal is calculated, and the noise component is proportionally subtracted from each channel to remove common-mode noise.
[0027] During the process of removing bad sectors, the corresponding columns of the fiber optic rolls in the DAS data after common-mode noise is removed by constructing a Boolean mask.
[0028] During the filtering process, bandpass filtering is applied to the DAS data after removing bad sectors to highlight the vehicle's trajectory signal.
[0029] In this embodiment, the vehicle trajectory signal enhancement of the feature image specifically includes: Perform absolute value processing on the feature image; The feature image after absolute value processing is normalized by interval summation.
[0030] With this configuration, this embodiment can enhance the vehicle trajectory signal. The feature map after the vehicle trajectory signal enhancement is as follows: Figure 2 As shown.
[0031] During the absolute value processing, the feature image is subjected to absolute value filtering. The energy threshold is determined by sorting the absolute values within the column, retaining only significant signals, suppressing low-amplitude noise, and keeping the top 10% of data with the largest absolute value in each column, while setting the rest to zero.
[0032] In the process of interval summation and normalization, the overall intensity characteristics of the signal within different time windows are extracted by summing by row intervals (windows) and normalizing the column sum of each window by the maximum absolute value, thereby eliminating the difference in dimensions and enhancing the vehicle trajectory signal.
[0033] like Figure 3 As shown, in this embodiment, vehicle trajectory signals in different directions are separated from the enhanced feature image to obtain a vehicle trajectory separation feature map, specifically including: The enhanced feature image data is transformed from the spatiotemporal domain to the frequency-wavenumber domain by Fourier transform, and the logarithmic amplitude spectrum and phase spectrum of the feature image are extracted. In the frequency domain-wavenumber domain, the logarithmic amplitude spectrum and phase spectrum of the feature image are subjected to sector filtering according to polar coordinate angles in different preset directions; The filtered logarithmic amplitude spectrum and phase spectrum are converted from the frequency domain to the wavenumber domain and back to the spatiotemporal domain by inverse Fourier transform, and the data with corresponding angles are reconstructed to obtain vehicle trajectory separation feature maps in different directions.
[0034] This embodiment filters and retains vehicle trajectory signals to suppress noise signals from other angles.
[0035] In this embodiment, the polar coordinate angles are 0-30° and 180°-210°, or 150°-180° and 330°-360°.
[0036] like Figure 4 As shown, in this embodiment, edge detection and Hough transform are performed on the segment feature maps respectively to obtain the coordinates of the straight line segments of the vehicle trajectory of each vehicle in each segment, specifically including: The Canny algorithm is used to separate the edges of vehicle trajectories in the feature maps of each segment, generating binarized edge images. By using Hough transform, the vehicle trajectory line segments are extracted from the binarized edge image of each segment, and the coordinates of the vehicle trajectory line segments of each vehicle in each segment are obtained.
[0037] In this embodiment, the average speed of all vehicles in each segment is obtained based on the endpoint coordinates of the straight line segments of each vehicle's trajectory within each segment. Specifically, this includes: The slope of each straight line segment is calculated based on the endpoint coordinates of the straight line segments of each vehicle trajectory within each segment. Calculate the average slope of each straight segment and use it as the average speed of all vehicles in each segment.
[0038] In this embodiment, the dynamic baseline threshold is the average speed of the vehicle within a first preset time period in the segment.
[0039] In this embodiment, based on the average speed of all vehicles in each segment and the dynamic baseline threshold of that segment, it is determined whether a low-speed event exists in that segment, specifically including: If the average speed of a certain segment is consistently lower than 85% of the dynamic baseline threshold during the second preset time period, it is determined that there is a low-speed event in that segment; otherwise, it is determined that there is no low-speed event in that segment.
[0040] It should be understood that the sum of the first preset time period and the second preset time period is less than or equal to the total travel time of the vehicle in the segment, and the first preset time period is earlier than the second preset time period in terms of time sequence, and the duration of the first preset time period is greater than the duration of the second preset time period.
[0041] With this configuration, this embodiment significantly reduces the false negatives and false positives of low-speed events caused by traffic flow fluctuations on different road segments by updating the dynamic baseline threshold in real time. In one specific embodiment, the first preset time period is the first 30 minutes of the segment, and the second preset time period is the last 6 minutes of the segment.
[0042] In one specific embodiment, the proposed highway low-speed event monitoring method based on DAS and dynamic thresholds was used to monitor a 15km long road segment for 3 months. The monitoring results included: 2 / 2 of the accidents occupying 3 lanes were identified, and the identification rate for accidents occupying 3 or more lanes was 100%; 4 / 4 of the accidents occupying 2 lanes were identified, and the identification rate for accidents occupying 2 or more lanes was 100%; 2 / 5 of the accidents occupying 1 lane were identified, and the identification rate for accidents occupying 1 or more lanes was 72.73%; 2 false alarms were reported. Within the monitoring area, the notification time of the highway low-speed event monitoring method based on DAS and dynamic thresholds is on average 6 minutes earlier than the manual notification time (it usually takes 15-20 minutes from the occurrence of an event to the management office receiving the news, but after being put into the system, the management office will receive the news 6-10 minutes after the event occurs), which speeds up the handling of accidents and reduces losses.
[0043] Similarly, a fixed threshold algorithm (with a low threshold) was used to monitor the 15km section of road for 3 months. The results were: 1 accident that occupied 3 lanes was missed, 2 accidents that occupied 2 lanes were missed, and 5 accidents that occupied 1 lane were missed.
[0044] A fixed threshold algorithm (with a high threshold) was used to monitor this 15km section of road for 3 months. The results were as follows: 2 / 2 of the accidents occupied 3 lanes, 4 / 4 of the accidents occupied 2 lanes, 2 / 5 of the accidents occupied 1 lane, and 5 false alarms.
[0045] This invention can effectively reduce the false alarm rate and false alarm rate of low-speed events.
[0046] Secondly, this invention also proposes a highway low-speed event monitoring system based on DAS and dynamic thresholds, comprising: The acquisition module is used to acquire the DAS signal of the target highway segment; The processing module is used to preprocess the DAS signal to obtain a feature image including vehicle trajectory signals; enhance the vehicle trajectory signals in the feature image; separate the vehicle trajectory signals in different directions in the enhanced feature image to obtain vehicle trajectory separation feature maps in different directions; divide the vehicle trajectory separation feature maps in different directions into segments according to the segment settings of the target highway segment to obtain multiple segment feature maps; perform edge detection and Hough transform on the segment feature maps to obtain the coordinates of the straight line segments of the vehicle trajectory of each vehicle in each segment; obtain the average speed of all vehicles in each segment based on the endpoint coordinates of the straight line segments of the vehicle trajectory in each segment; determine whether there is a low-speed event in the segment based on the average speed of all vehicles in each segment and the dynamic baseline threshold of the segment; if so, issue an alarm.
[0047] In this embodiment, the system also includes an optical cable pre-laid on the target highway section and a DAS system connected to the optical cable. The DAS system is communicatively connected to the acquisition module.
[0048] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.
Claims
1. A method for monitoring low-speed events on highways based on DAS and dynamic thresholds, characterized in that, include: Acquire DAS signals for the target highway segment; The DAS signal is preprocessed to obtain a feature image including vehicle trajectory signal; Enhance vehicle trajectory signals in feature images; Vehicle trajectory signals in different directions are separated in the enhanced feature image to obtain vehicle trajectory separation feature maps in different directions; Based on the segmentation of the target highway segment, the vehicle trajectory separation feature map for different directions is divided into segments to obtain multiple segment feature maps; Edge detection and Hough transform are performed on the feature maps of each segment to obtain the coordinates of the straight line segments of the vehicle trajectory of each vehicle in each segment. The average speed of all vehicles in each segment is obtained by using the endpoint coordinates of the straight line segments of each vehicle's trajectory within each segment. Based on the average speed of all vehicles in each segment and the dynamic baseline threshold of that segment, determine whether a low-speed event exists in that segment. If so, then an alarm will be triggered.
2. The method for monitoring low-speed events on highways based on DAS and dynamic thresholds according to claim 1, characterized in that, Preprocessing includes: Common-mode noise removal, bad sector removal, and filtering.
3. The method for monitoring low-speed events on highways based on DAS and dynamic thresholds according to claim 1, characterized in that, Enhancement of vehicle trajectory signals in feature images, specifically including: Perform absolute value processing on the feature image; The feature image after absolute value processing is normalized by interval summation.
4. The method for monitoring low-speed events on highways based on DAS and dynamic thresholds according to claim 1, characterized in that, The vehicle trajectory signals in different directions are separated from the enhanced feature image to obtain a vehicle trajectory separation feature map, specifically including: The enhanced feature image data is transformed from the spatiotemporal domain to the frequency-wavenumber domain by Fourier transform, and the logarithmic amplitude spectrum and phase spectrum of the feature image are extracted. In the frequency domain-wavenumber domain, the logarithmic amplitude spectrum and phase spectrum of the feature image are subjected to sector filtering according to polar coordinate angles in different preset directions; The filtered logarithmic amplitude spectrum and phase spectrum are converted from the frequency domain to the wavenumber domain and back to the spatiotemporal domain by inverse Fourier transform, and the data with corresponding angles are reconstructed to obtain vehicle trajectory separation feature maps in different directions.
5. The method for monitoring low-speed events on highways based on DAS and dynamic thresholds according to claim 4, characterized in that, The polar coordinate angles are 0-30° and 180°-210°, or 150°-180° and 330°-360°.
6. The method for monitoring low-speed events on highways based on DAS and dynamic thresholds according to claim 1, characterized in that, Edge detection and Hough transform are performed on the feature maps of each segment to obtain the coordinates of the straight line segments of the vehicle trajectory for each vehicle in each segment, specifically including: The Canny algorithm is used to separate the edges of vehicle trajectories in the feature maps of each segment, resulting in a binarized edge image for each segment. By using Hough transform, the vehicle trajectory line segments are extracted from the binarized edge image of each segment, and the coordinates of the vehicle trajectory line segments of each vehicle in each segment are obtained.
7. The method for monitoring low-speed events on highways based on DAS and dynamic thresholds according to claim 1, characterized in that, Based on the endpoint coordinates of the straight segments of each vehicle's trajectory within each segment, the average speed of all vehicles within each segment is obtained, specifically including: The slope of each straight line segment is calculated based on the endpoint coordinates of the straight line segments of each vehicle trajectory within each segment. Calculate the average slope of each straight segment and use it as the average speed of all vehicles in each segment.
8. The method for monitoring low-speed events on highways based on DAS and dynamic thresholds according to claim 1, characterized in that, The dynamic baseline threshold is the average speed of the vehicle within the first preset time period of the segment; Specifically, based on the average speed of all vehicles within each segment and the dynamic baseline threshold of that segment, it is determined whether a low-speed event exists in that segment, including: If the average speed of a certain segment is consistently lower than 85% of the dynamic baseline threshold during the second preset time period, it is determined that a low-speed event exists in that segment; otherwise, it is determined that no low-speed event exists in that segment. Wherein, the sum of the first preset time period and the second preset time period is less than or equal to the total driving time of the vehicle in the section, and the first preset time period is earlier than the second preset time period in terms of time sequence, and the duration of the first preset time period is greater than the duration of the second preset time period.
9. A highway low-speed event monitoring system based on DAS and dynamic threshold, characterized in that, include: The acquisition module is used to acquire the DAS signal of the target highway segment; The processing module is used to preprocess the DAS signal to obtain a feature image including the vehicle trajectory signal; The process involves enhancing vehicle trajectory signals in the feature image; separating vehicle trajectory signals in different directions within the enhanced feature image to obtain vehicle trajectory separation feature maps in different directions; dividing the vehicle trajectory separation feature maps in different directions into segments according to the segment settings of the target highway segment to obtain multiple segment feature maps; performing edge detection and Hough transform on each segment feature map to obtain the coordinates of the straight line segments of each vehicle's trajectory in each segment; obtaining the average speed of all vehicles in each segment based on the endpoint coordinates of the straight line segments of each vehicle's trajectory in each segment; and determining whether a low-speed event exists in each segment based on the average speed of all vehicles in each segment and the dynamic baseline threshold of that segment. If so, then an alarm will be triggered.
10. The highway low-speed event monitoring system based on DAS and dynamic threshold as described in claim 9, characterized in that, It also includes optical cables pre-laid on the target highway section and a DAS system connected to the optical cables, with the DAS system communicating with the acquisition module.