A moving car detection method based on distributed optical fiber short-line spatiotemporal information

By setting the acquisition point on the optical fiber vibration sensing unit, calculating short-term energy and drawing a space-time scatter plot, and using fitted straight lines to judge the car's driving, the problem of difficulty in car driving recognition in distributed fiber intrusion detection is solved, and efficient and low-cost car detection is achieved.

CN118397851BActive Publication Date: 2025-08-26ZHEJIANG UNIV +1
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
CN202410448430.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-04-15
Publication Date
2025-08-26
Estimated Expiration
2044-04-15

AI Technical Summary

Technical Problem

The existing distributed fiber intrusion detection method is difficult to quickly and accurately identify the vehicle's driving when the single point original signal caused by the car is highly consistent with other intrusion signals, resulting in low detection efficiency.

Method used

By laying the optical fiber vibration sensing unit along the direction of the car, setting the acquisition point, collecting and calculating short-term energy in real time, drawing a space-time scatter plot, using the inclination angle of the fitted line to determine whether a car has passed by, and filtering the interference signals in combination with buffering and energy thresholds to achieve specific categories of car detection.

Benefits of technology

Effectively eliminates interference signals, improves the accuracy and efficiency of automobile inspection, reduces system cost and energy consumption, is suitable for large-scale monitoring, and is easy to integrate with other detection methods.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118397851B_ABST
    Figure CN118397851B_ABST
Patent Text Reader

Abstract

The present invention proposes a method for detecting moving cars based on distributed optical fiber short-line spatiotemporal information. First, optical fiber vibration sensing units are laid along the direction of vehicle travel to form a distributed optical fiber sensing system. The one-second short-term energy of each collection point on the distributed optical fiber is calculated, and the geographic information is saved and cached for a period of time. The point to be detected is input, and based on its geographic information, the short-term energy data of the collection points cached within a certain distance around the point to be detected are counted. An energy threshold is set, and the spatiotemporal information of the data points exceeding the threshold is used to create a scatter plot. A straight line is fitted based on the least squares method, and when the inclination of the straight line is within a certain range, it is determined that a car has passed. The present invention utilizes the physical characteristics of vehicle travel and fully combines the spatiotemporal information of distributed optical fibers. It can be judged using only the short-term energy characteristics of the original data, thereby improving the real-time nature of the judgment and the detection efficiency. At the same time, it can be used in combination with other single-point-based detection strategies and has a wide range of applicability.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of distributed optical fiber vibration perimeter intrusion detection and security, and more specifically, relates to a moving car detection method based on distributed optical fiber short-line spatiotemporal information. Background Art

[0002] Distributed fiber optic sensing is based on the one-dimensional spatial characteristics of optical fiber. Its sensing principle treats the physical parameter to be measured as a function of the fiber's position and length. By continuously measuring the external parameters along the fiber in real time, the spatial distribution of the measured physical quantity and its changes over time can be obtained. The entire system is based on a single sensing fiber. Each segment of the fiber can be considered a sensing device and also serves as a channel for other sensing devices to transmit sensing information. Because the entire system is built on a single fiber, its cost is greatly reduced. Monitoring systems based on distributed fiber optics have the advantages of high sensitivity, wide coverage, and accurate positioning, and are being adopted by an increasing number of monitoring systems.

[0003] Distributed fiber-optic intrusion detection methods are a key factor in the development of distributed intrusion detection technology. Current distributed intrusion detection methods are mostly based on a single detection point, determining the nature of the intrusion based on the disturbance signal detected at that point. Pattern recognition of the detection signal has always been a key and challenging aspect of distributed fiber-optic detection. Currently, mainstream identification methods use audio analogies to intrusion signals, leveraging audio features for classification and identification. Most perimeter intrusion detection facilities are located adjacent to roads. Vehicles traveling along these roads can easily interfere with the intrusion system, generating signals with audio characteristics similar to those of the intrusion signal. Since the single-point interference signal is highly consistent with the audio characteristics of the type being detected, this fundamentally complicates mainstream detection methods. Therefore, it is crucial to quickly and accurately identify vehicles in motion and eliminate interference when applying distributed fiber-optic intrusion detection methods. Summary of the Invention

[0004] In order to solve the problem that existing distributed optical fiber intrusion detection methods ignore the spatiotemporal characteristics of specific categories and become inefficient when the single-point original signal caused by vehicle movement is too consistent with other intrusion detection categories, the present invention proposes a moving vehicle detection method based on distributed optical fiber short-line spatiotemporal information.

[0005] The technical solution adopted by the present invention is that the method for detecting a moving vehicle based on distributed optical fiber short-line spatiotemporal information comprises:

[0006] 1) Laying optical fiber vibration sensing units along the direction of vehicle travel; k collection points are evenly set along the laying direction of the optical fiber vibration sensing units, and the collection points collect raw data in real time. A short-term cycle is set, and the short-term energy of each collection point is calculated based on the raw data in each short-term cycle and cached; k is a natural number greater than 1;

[0007] 2) According to the location of the point to be detected, the part of the cached short-time energy corresponding to the vicinity of the point to be detected is retrieved;

[0008] 3) Set an energy threshold, compare the retrieved short-time energy with the energy threshold, screen out the short-time energy that exceeds the energy threshold, and draw a spatiotemporal scatter plot based on the collection point location and collection time corresponding to the screened short-time energy;

[0009] 4) Use the points of the time-space scatter plot to fit a straight line, calculate the inclination of the fitted straight line, and judge whether a car passes by the point to be detected based on the size of the inclination.

[0010] Furthermore, in step 1), the distance x between adjacent collection points on the optical fiber vibration sensing unit is less than 30 m.

[0011] Furthermore, in step 1), the raw data collected by the collection points in real time are specifically: each collection point performs sr sampling times per second, and the raw data obtained by the jth collection point at the rth sampling time in a short period is d rj ;

[0012] The short-time period is set to 1 second; the short-time energy of each acquisition point is calculated based on the original data in each short-time period, specifically:

[0013]

[0014] Among them, e j Represents the short-time energy of the jth acquisition point in a short-time period.

[0015] Furthermore, in step 1), the cache is specifically as follows: establishing a cache area, storing the short-time energy data of the entire line in the most recent second into the cache area every second, and not discarding the data if the cache area is not saturated; if the cache area is saturated, discarding the earliest short-time energy data of the entire line in the cache area in the first second.

[0016] Furthermore, in step 2), the buffer area is:

[0017] E kt =(e ij ) k×t (i=1,2,...,t; j=1,2,,...,k)

[0018] Among them, e ijIt represents the data point in the buffer area, that is, the short-time energy of the jth collection point in the i-th short-time period, t is the length of the time period cached in the buffer area, and k is the number of collection points along the entire line.

[0019] Furthermore, in the step 2),

[0020] The vicinity of the point to be detected, that is, the position range corresponding to the retrieved data is Where j represents the position of the point to be detected, d = V max t represents the range size, V max represents the maximum speed limit of the car on this road section, and t represents the length of the time period corresponding to the cached data in step 1).

[0021] Furthermore, in step 3), the energy threshold is the average value of the short-time energy of all the collection points within a short period when no car passes by and no intrusion event occurs, which is used to filter out normal background noise data.

[0022] Furthermore, in step 3), the specific method of drawing the spatiotemporal scatter diagram according to the collection point positions and collection times corresponding to the screened short-time energy is as follows: ij , where i represents the relative position of the corresponding collection point, j represents the corresponding collection time, and a spatiotemporal scatter plot is drawn with i as the horizontal axis and j as the vertical axis;

[0023] If there are multiple points with the same horizontal coordinate in the space-time scatter diagram, the point with the largest short-time energy value is retained and the other points are deleted.

[0024] Furthermore, in step 4), the fitting method is the least square method, and the linear equation obtained by fitting is: y=ax+b, then the magnitude of the inclination angle is θ=arctan a;

[0025] The method of judging whether a car has passed by the detection point according to the size of the inclination angle is as follows: if the size of the inclination angle is less than the set angle, it is judged that a car has passed by the detection point; otherwise, it is judged that no car has passed by the detection point; the set angle is Where V min is the minimum speed of the car when the vibration information generated by the car can be detected by the optical fiber vibration sensing unit, and x0 represents the distance between adjacent collection points on the optical fiber vibration sensing unit.

[0026] The present invention further proposes a moving vehicle detection system based on distributed optical fiber short-line spatiotemporal information, for implementing the above method, the system comprising:

[0027] The acquisition and calculation module is used to acquire raw data in real time and calculate short-time energy based on the raw data;

[0028] The cache retrieval module is used to cache the full-line short-time energy data of all frames in the recent certain time period, and retrieve the short-line short-time energy data according to the geographical information of the point to be detected;

[0029] A plotting module, used to filter data points based on energy thresholds and draw spatiotemporal scatter plots;

[0030] The fitting detection module is used to fit a straight line according to the spatiotemporal scatter plot and determine whether a car passes by the point to be detected based on the inclination of the fitted straight line.

[0031] In summary, the technical solution conceived by the present invention has the following advantages compared with the prior art:

[0032] (1) The present invention takes into account the physical characteristics of automobile driving and fully utilizes the temporal and spatial information of distributed optical fiber signals, thereby eliminating the interference of other single-point events on automobile driving judgment.

[0033] (2) The present invention stores data in the form of short-term energy per second, extracting signal features to the greatest extent possible and significantly reducing space requirements. Furthermore, the method for determining vehicle usage is simple and effective, and offers superior performance compared to traditional machine learning or deep learning methods.

[0034] (3) The present invention uses a fiber optic sensing system as a detection element. Compared with camera detection methods, it has lower cost and lower energy consumption, is more suitable for large-scale monitoring, and is easier to maintain in harsh environments. At the same time, when undertaking more complex detection tasks, it is also easy to integrate with methods based on other detection equipment such as radar and cameras.

[0035] (4) Although the present invention is only targeted at a specific category of vehicle driving detection, it can be well integrated with other detection methods based on fiber optic sensing systems, such as intrusion event detection methods based on machine learning or deep learning. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 This is a flow chart of a moving vehicle detection method based on distributed optical fiber short-line spatiotemporal information provided by an embodiment of the present invention.

[0037] Figure 2 It is the raw data per second of a certain collection point provided by the embodiment of the present invention.

[0038] Figure 3 This is a short-time energy graph provided by an embodiment of the present invention when a car passes by a certain collection point.

[0039] Figure 4 This is a short-time energy diagram provided by an embodiment of the present invention when a human intrusion action occurs at a certain collection point.

[0040] Figure 5 、6 This is a time-space scatter plot comparison diagram of a car passing by and a human intrusion action at a certain collection point after background noise screening, provided by an embodiment of the present invention.

[0041] Figure 7 This is a spatiotemporal scatter plot of a certain collection point where a car passes by after processing provided by an embodiment of the present invention.

[0042] Figure 8 This is a spatiotemporal scatter diagram of a certain collection point having a human intrusion action after processing provided by an embodiment of the present invention.

[0043] Figure 9 It is a straight line fitted to a spatiotemporal scatter plot of a certain collection point where a car passes by, provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0044] The present invention will be further described and illustrated below in conjunction with specific embodiments. The embodiments are merely illustrative of the present disclosure and do not limit its scope. The technical features of the various embodiments of the present invention may be combined accordingly, provided that there is no conflict between them.

[0045] To make the objectives, technical solutions, and advantages of the present invention more clearly understood, the following detailed description of the method for detecting moving vehicles based on distributed optical fiber short-line spatiotemporal information is provided in conjunction with the accompanying drawings and embodiments. It should be noted that the specific embodiments described herein are intended only to illustrate the present invention and are not intended to limit the present invention.

[0046] In the embodiment of the present invention, a moving car detection method based on distributed optical fiber short-line spatiotemporal information is provided. Figure 1 Shown, including:

[0047] Step 1: Fiber optic vibration sensing units are laid out along the direction of vehicle travel to form a distributed fiber optic sensing system. Based on the raw data collected per second at each collection point, the short-term energy per second at each collection point is calculated and cached, retaining geographic information. The raw data is vibration information directly collected by the fiber optic vibration sensing units at each collection point.

[0048] When laying the optical fiber vibration sensing unit, it must be laid on the ground within 5m from the vehicle's driving route.

[0049] Step 2: The buffer only stores the short-term energy data of the entire line for a period of time. When the buffer is saturated, the earliest one-second short-term energy data of the entire line is discarded to store the latest full-line information. Based on the geographic information of the input point to be tested, the short-term energy data cached within a certain range near the point to be tested is retrieved;

[0050] Step 3: Set an appropriate energy threshold, compare the retrieved short-term energy data with the threshold, obtain the spatiotemporal information of the data points that exceed the energy threshold, and plot the spatiotemporal points into a spatiotemporal scatter plot with time as the horizontal axis and space as the vertical axis;

[0051] Step 4: Fit a straight line using the points in the spatiotemporal scatter plot. Calculate the inclination of the fitted line with respect to the horizontal axis. If the inclination is within a certain range, it is considered that a car has passed by.

[0052] Step 1 includes:

[0053] Step 1-1 The original signal of the entire line for one second can be expressed as D sr×k =(d rj ) sr×k (r=1,2,…,sr;j=1,2,…,k) where k is the geographical dimension of the sampling point, i.e. the number of sampling points along the entire line, sr is the sampling rate of the distributed optical fiber system, and the raw data collected per second at a certain sampling point is as follows: Figure 2 As shown, the sampling rate sr = 1600 Hz is used in this embodiment. Generally speaking, the sampling rate of distributed fiber optic sensing systems ranges from 1 kHz to 10 kHz, and storing the entire line signal takes up too much space. Since a car passing around the fiber causes fiber vibrations, resulting in changes in vibration energy, only one second of energy at each sampling point needs to be cached to retain the vibration information at that moment, while also saving space.

[0054] The energy value of the original data is calculated along the frame direction according to each geographical location to obtain a buffer containing spatiotemporal information. The short-term energy is calculated as follows:

[0055]

[0056] Step 1-2 caches the energy values ​​of all collection points by seconds. The cache area containing short-term energy within a period of time can be expressed as: E kt =(e ij ) k×t (i=1,2,...,k; j=1,2,,...,t), e ij is the data point stored in the buffer area, that is, the short-time energy of the j-th frame of the i-th collection point. i and j are the time information and geographic information of the short-time energy, respectively, which together serve as the spatiotemporal information of the short-time energy. t is the length of the time period cached in the buffer area, which is determined according to the number of collection points along the entire line, the real-time performance required for detection, and the system performance.

[0057] The specific implementation method for determining the caching time range and the retrieved geographical distance range is to estimate the normal driving speed of the vehicle in the distributed optical fiber application scenario, where 30 km / h < v < 100 km / h. According to 1 m / s = 3.6 km / h, the driving speed of the vehicle can be obtained as 8.3 m / s < v < 27.8 m / s. According to the number of acquisition points k along the whole line, the real-time requirement for detection, and the system performance, the caching time range t can be determined first, and then the geographical range to be retrieved can be obtained according to d = v·t as follows: j is the geographical location of the monitoring point. Specifically, in this embodiment, the caching time length is 20 s, and the geographical distance range is 200 m. The short-term energy diagram of the optical fiber when the vehicle passes through a certain acquisition point is as shown in Figure 3 shown, and the short-term energy diagram of the human intrusion action is as shown in Figure 4 shown. It can be seen from the energy diagram that the influence of the vehicle driving on the optical fiber shows obvious regularity compared with the human intrusion action. In the embodiment, an acquisition point is set every 4 m of the actual geographical distance, which is a defense zone coordinate. Therefore, the actual position is replaced by the defense zone coordinate, which is the quantity on the horizontal axis in the schematic diagram of the embodiment.

[0058] Further, the specific implementation of drawing the spatio-temporal scatter plot in step 3 is as follows:

[0059] According to the data format E of the buffer kt , a threshold E0 is selected. If e ij > E0, the time and geographical information i, j of this data point are saved. This step is to filter out the normal background noise data. The threshold E0 can be set by statistically analyzing the data in the normal environment and is used to filter out the background noise generated by the environment, pedestrians, wild animals, bicycles, etc. In this embodiment, E0 = 7000 is selected. The scatter plot generated after filtering the background noise is as shown in Figure 5 , and for comparison, Figure 6 is the scatter plot generated by the human intrusion action. Traverse the buffer E kt , and generate a sequence containing the geographical and time information of all points exceeding the energy threshold: {(j1, i1), (j2, i2), (j3, i3),...}. Taking the geographical dimension as the horizontal axis and the time dimension as the vertical axis, the spatio-temporal scatter plot can be drawn, as shown in Figure 7 , and for comparison, Figure 8 is the scatter plot generated by the human intrusion action. According to Figure 7 and Figure 8 comparison, it can be seen that the spatio-temporal scatter plot of the vehicle driving has obvious distribution rules. Assume that the abscissa j in the scatter plot n corresponds to multiple ordinates i1, i2, i3.... According to the rule:

[0060]

[0061] The points that can be left by filtering (i n ,j n ). After screening, the short-time energy scattered point set is obtained:

[0062] {(j1,i1),(j2,i2),(j3,i3),...,(j m ,i m )}

[0063] Furthermore, the specific implementation method of calculating the inclination angle between the fitting straight line and the horizontal axis in step 4 is:

[0064] Assuming that the equation of the fitted line is: y = ax + b, the inclination angle can be obtained as θ = arctan a

[0065] The specific implementation method of the fitting straight line is:

[0066] The fitting method is implemented using the least squares method, and the fitting formula is as follows:

[0067]

[0068] The fitting straight line is obtained by the fitting method. The fitting straight line diagram obtained in the embodiment is as follows Figure 6 If the angle θ of the fitting line is less than the set angle, it is determined that a car has passed through the defense zone. The set angle is Where V min is the minimum speed of the car when the vibration information generated by the car can be detected by the optical fiber vibration sensing unit, x0 represents the distance between adjacent collection points on the optical fiber vibration sensing unit; V min Specifically, in the embodiment, the angle θ of the fitting line is 11°, and the set angle is 5°, so it is determined that a car is passing by.

[0069] Compared to existing technologies, the method in this embodiment fully utilizes the temporal and spatial information available through distributed optical fiber. Because the temporal and spatial characteristics of a passing vehicle are distinct from those of other events, this method effectively avoids interference from other environmental factors or intrusions, making it suitable for complex environments with inclement weather. Compared to single-point detection methods, this method significantly improves detection performance and is easily integrated with other detection methods.

[0070] The accompanying drawings illustrating the embodiments of the present invention serve to more clearly illustrate the objectives, technical solutions, and advantages of the present invention. It should be noted that the specific embodiments described herein are intended only to illustrate the present invention and are not intended to limit the present invention. Any equivalent substitutions, modifications, and the like made within the methodologies and principles provided by the present invention are intended to be included within the scope of protection of the present invention.

Claims

1. A moving car detection method based on distributed optical fiber short-line spatiotemporal information, characterized in that: include: 1) Fiber optic vibration sensing units are laid along the direction of vehicle travel. K collection points are evenly set along the laying direction of the fiber optic vibration sensing units. The collection points collect raw data in real time. A short-term cycle is set. In each short-term cycle, the short-term energy of each collection point is calculated based on the raw data and cached. k is a natural number greater than 1; 2) According to the location of the point to be detected, the part of the cached short-time energy corresponding to the area near the point to be detected is retrieved; 3) Set an energy threshold, compare the retrieved short-time energy with the energy threshold, filter out the short-time energy that exceeds the energy threshold, and draw a spatiotemporal scatter plot based on the collection point location and collection time corresponding to the filtered short-time energy; 4) Fit a straight line using the points in the spatiotemporal scatter plot, calculate the inclination of the fitted straight line, and determine whether a car has passed by the point to be detected based on the size of the inclination; In step 1), the raw data collected by the collection points in real time are specifically: each collection point performs sr sampling times per second, and the raw data obtained by the jth collection point at the rth sampling time in a short time period is ; The short-time period is set to 1 second; the short-time energy of each acquisition point is calculated based on the original data in each short-time period, specifically: ; in, Represents the short-time energy of the jth acquisition point in a short-time period; In step 1), the cache is specifically as follows: establishing a cache area, storing the short-term energy data of the entire line in the most recent second into the cache area every second, and not discarding the data if the cache area is not saturated; if the cache area is saturated, discarding the short-term energy data of the entire line in the earliest second in the cache area; The buffer area is: ; in, represents the data point in the buffer area, that is, the short-time energy of the jth collection point in the i-th short-time period, t is the length of the time period cached in the buffer area, and k is the number of collection points along the entire line; In the step 2); The vicinity of the point to be detected, that is, the position range corresponding to the retrieved data is , where j represents the position of the point to be detected, d = V max t represents the range size, V max represents the maximum speed limit of the car on the road section, and t represents the length of the time period corresponding to the cached data in step 1); In step 3), the energy threshold is the average value of the short-term energy of all the collection points in a short period when no car passes by and no intrusion event occurs, which is used to filter out normal background noise data; In step 3), the specific method of drawing the spatiotemporal scatter diagram based on the collection point positions and collection times corresponding to the short-time energy screened out is: , where i represents the relative position of the corresponding collection point, j represents the corresponding collection time, and a spatiotemporal scatter plot is drawn with i as the horizontal axis and j as the vertical axis; If there are multiple points with the same horizontal coordinate in the space-time scatter diagram, the point with the largest short-time energy value is retained and the other points are deleted; In step 4), the fitting method is the least squares method, and the linear equation obtained by fitting is: , then the magnitude of the inclination angle is ; The method of judging whether a car has passed by the detection point according to the size of the inclination angle is as follows: if the size of the inclination angle is less than the set angle, it is judged that a car has passed by the detection point; otherwise, it is judged that no car has passed by the detection point; the set angle is ,in is the minimum speed of the car when the vibration information generated by the car can be detected by the optical fiber vibration sensing unit, and x0 represents the distance between adjacent collection points on the optical fiber vibration sensing unit.

2. The method for detecting a moving vehicle based on distributed optical fiber short-line spatiotemporal information according to claim 1, characterized in that: In the step 1), the distance x between adjacent collection points on the optical fiber vibration sensing unit is less than 30 m.

3. A moving vehicle detection system based on distributed optical fiber short-line spatiotemporal information, used to implement the method according to claim 1, characterized in that: The system comprises: The acquisition and calculation module is used to acquire raw data in real time and calculate short-time energy based on the raw data; The cache retrieval module is used to cache the full-line short-time energy data of all frames in the recent certain time period, and retrieve the short-line short-time energy data according to the geographical information of the point to be detected; A plotting module, used to filter data points based on energy thresholds and draw spatiotemporal scatter plots; The fitting detection module is used to fit a straight line according to the spatiotemporal scatter plot and determine whether a car passes by the point to be detected based on the inclination of the fitted straight line.

Citation Information

Patent Citations

  • Intrusion detecting and positioning method for distributed optical fiber fence

    CN102360519A