A full-time and all-weather road perception method and system based on vibration sensing

Abstract

The present invention relates to a method and system for all-weather road perception in all time and space domains based on vibration sensing, wherein the method obtains vibration sensing data that is less affected by bad weather and other roadside detection data that is easily affected by weather, corrects and synchronizes other roadside detection data based on the time and space information of the vibration sensing data, and fuses the various roadside detection data according to the synchronized spatial information obtained after synchronization, so as to achieve full-scene perception in all weather conditions. Compared with the prior art, the present invention performs synchronous calibration on various roadside detection data based on vibration sensing data, reduces the impact of bad weather on the accuracy of sensing data, and solves the problem in the prior art that the perception of the situation of the highway is seriously affected by weather. At the same time, the present invention also realizes accurate perception of the highway in all time and space domains by fusing vibration sensing data and roadside detection data and utilizing the characteristics of two different detection methods.

Description

Technical Field

[0001] The present invention relates to the technical field of traffic monitoring, and in particular to a vibration sensing-based full-time and space-domain all-weather road perception method and system. Background Art

[0002] At present, the equipment relied on for information collection on highways is mainly video detectors, laser radars and millimeter wave radars. Among them, video detectors can provide visual information for traffic accident management. They have the advantages of wide detection range, easy mobility and low maintenance cost. However, video detectors are easily affected by external environment, such as poor external lighting conditions, rainy and foggy weather, backlight conditions, etc. Under these conditions, the false detection rate of video detectors will be greatly increased.

[0003] Similarly, although lidar and millimeter-wave radar have advantages such as high resolution and strong anti-interference ability, their high cost and severe attenuation of detection capabilities in adverse weather conditions still make it impossible to achieve full-time and space-time collection of highway traffic information.

[0004] In short, although the existing highway information collection technology is mature, its perception of the highway is greatly affected by the climate. Under the harsh weather conditions of the highway, the roadside identification and positioning accuracy are low, and the detection time and spatial range are limited, making it impossible to accurately perceive the highway conditions around the clock. Summary of the invention

[0005] In view of this, it is necessary to provide a full-time and space-domain all-weather road perception method and system based on vibration sensing to solve the problem in the prior art that the perception of highway conditions is seriously affected by weather and cannot accurately perceive the highway conditions around the clock.

[0006] This patent solves the problems of low roadside identification and positioning accuracy, limited detection time and spatial range under harsh weather conditions on highways.

[0007] In order to achieve the above technical objectives, the present invention adopts the following technical solutions:

[0008] In a first aspect, the present invention provides a vibration sensing-based all-time and all-weather road perception method, comprising:

[0009] Acquiring vibration sensing data and a plurality of roadside detection data, wherein the vibration sensing data and the roadside detection data both include temporal and spatial information;

[0010] Based on the spatiotemporal information of the vibration sensing data, the spatiotemporal information of the roadside detection data is corrected and synchronized to obtain the synchronized spatial information corresponding to the roadside detection data;

[0011] Taking the spatiotemporal information of the vibration sensing data and the synchronization information data as a reference, the vibration sensing data and the roadside detection data are integrated to obtain a multi-dimensional information data set corresponding to each monitored vehicle;

[0012] Full scene perception is performed based on the multi-dimensional information data set to obtain event perception results.

[0013] Furthermore, the spatial reference coordinate system of the vibration sensor data is a kilometer pile positioning system, and the spatial reference coordinate systems of the multiple roadside detection data include the absolute coordinate system and the relative coordinate system; the time and space information of the roadside detection data is corrected and synchronized based on the time and space information of the vibration sensor data to obtain the synchronized spatial information corresponding to the road test detection data, including:

[0014] Performing time calibration on the time information in the vibration sensing data and the roadside detection data;

[0015] According to the kilometer pile positioning system and the absolute coordinate system, a correction mapping set is established, wherein the correction mapping set includes a mapping relationship between reference kilometer pile positioning coordinates under the kilometer pile positioning system and reference absolute coordinates under the absolute coordinate system;

[0016] According to the correction mapping set, the relative coordinates based on the relative coordinate system in the roadside detection data are synchronized and the synchronized spatial information is obtained, and the synchronized spatial information includes the reference kilometer pile positioning coordinates and the reference absolute coordinates.

[0017] Further, the kilometer pile positioning coordinates include a pile number and a lane number, and the absolute coordinates of the absolute coordinate system include longitude and latitude; and establishing a correction mapping set according to the kilometer pile positioning system and the absolute coordinate system includes:

[0018] According to the kilometer pile positioning system, a plurality of the reference kilometer pile positioning coordinates are obtained, and the actual lane area corresponding to each of the reference kilometer pile positioning coordinates is obtained;

[0019] Selecting an actual position object from each of the actual lane areas;

[0020] Based on the absolute coordinate system, obtaining the absolute coordinates of each of the actual position objects;

[0021] The absolute coordinates of the actual position object corresponding to each of the reference kilometer pile positioning coordinates are used as the reference absolute coordinates to obtain the correction mapping set.

[0022] Further, the spatial information based on the relative coordinate system in the roadside detection data includes relative coordinates in the relative coordinate system, and the relative coordinates include equipment installation position and relative position coordinates; and synchronizing the relative coordinates based on the relative coordinate system in the roadside detection data according to the correction mapping set and obtaining the synchronized spatial information includes:

[0023] According to the installation position of the equipment, the actual kilometer pile positioning coordinates corresponding to each relative coordinate in the kilometer pile positioning system are obtained;

[0024] According to the actual kilometer pile positioning coordinates, selecting the reference kilometer pile positioning coordinates having the smallest distance from the actual kilometer pile positioning coordinates from the correction mapping set;

[0025] According to the selected reference kilometer pile positioning coordinates and the corresponding reference absolute coordinates, the synchronous spatial information corresponding to the relative coordinates is obtained.

[0026] Furthermore, the spatiotemporal information of the vibration sensing data and the synchronous information data are used as a reference, the vibration sensing data and the roadside detection data are fused, and a multi-dimensional information data set corresponding to each monitored vehicle is obtained, including:

[0027] Establishing a multi-dimensional information data model based on the vibration sensing data and the drive test detection data;

[0028] Matching the time information of the roadside detection data and the corresponding synchronous spatial information with the time information and spatial information of the vibration sensor data respectively, and fusing the successfully matched roadside detection data and the vibration sensor data based on the multi-dimensional information data model to obtain the multi-dimensional information data set corresponding to each monitored vehicle;

[0029] According to the monitored vehicle corresponding to the multi-dimensional information data set, a unique identification ID is assigned to the corresponding monitored vehicle;

[0030] According to the unique identification ID, the data of the monitored vehicle is tracked and the corresponding multi-dimensional information data set is updated.

[0031] Furthermore, the roadside detection data includes relative detection data based on the relative coordinate system and absolute detection data based on the absolute coordinate system; the time information of the roadside detection data and the corresponding synchronous spatial information are matched with the time information and spatial information of the vibration sensor data respectively, and based on the multi-dimensional information data model, the successfully matched roadside detection data and the vibration sensor data are fused to obtain the multi-dimensional information data set corresponding to each monitored vehicle, including:

[0032] Matching the time information of the relative detection data and the corresponding synchronous spatial information with the time information and spatial information of the vibration sensor data respectively, and fusing the successfully matched relative detection data and the vibration sensor data based on the multi-dimensional information data model to obtain an initial multi-dimensional information data set corresponding to each monitored vehicle;

[0033] The time information and space information of the absolute detection data are matched with the time information and space information of the initial multidimensional information data set respectively, and based on the multidimensional information data model, the successfully matched absolute detection data and the initial alignment information data set are fused to obtain the multidimensional information data set corresponding to each monitored vehicle.

[0034] Furthermore, the time information of the relative detection data and the corresponding synchronous spatial information are matched with the time information and spatial information of the vibration sensor data respectively, and based on the multi-dimensional information data model, the successfully matched relative detection data and the vibration sensor data are fused to obtain an initial multi-dimensional information data set corresponding to each monitored vehicle, including:

[0035] Matching the time information of the relative detection data with the time information of the vibration sensing data to obtain time-matched relative data;

[0036] Matching the synchronous spatial information corresponding to the time matching relative data with the spatial information of the vibration sensing data to obtain time-space matching relative data;

[0037] Based on the multi-dimensional information data model, the time-space matching relative data and the vibration sensing data are integrated to obtain the initial multi-dimensional information data set;

[0038] The time information of the initial multi-dimensional information data set is the time information of the relative detection data, and the space information of the initial multi-dimensional information data set is the matching synchronous space information.

[0039] Furthermore, the time information and space information of the absolute detection data are matched with the time information and space information of the initial multidimensional information data set respectively, and based on the multidimensional information data model, the successfully matched absolute detection data and the initial alignment information data set are fused to obtain the multidimensional information data set corresponding to each monitored vehicle, including:

[0040] Matching the time information of the absolute detection data with the time information of the initial multi-dimensional information data set to obtain time-matched absolute data;

[0041] Matching the spatial information of the time-matched absolute data with the spatial information of the initial multi-dimensional information data set to obtain time-space matched absolute data;

[0042] Based on the multidimensional information data model, the spatiotemporal matching absolute data and the initial multidimensional information data set are fused to obtain the multidimensional information data set.

[0043] Furthermore, performing full scene perception according to the multi-dimensional information data set to obtain event perception results includes:

[0044] Construct a set of vehicle behavior patterns and obtain mathematical characteristics of each vehicle behavior pattern;

[0045] Obtaining a vehicle behavior mathematical representation model according to the vehicle behavior pattern set and the mathematical characteristics;

[0046] According to the multi-dimensional information data set, based on the vehicle behavior mathematical characterization model, the event perception result is obtained.

[0047] In a second aspect, the present invention further provides a full-time and space-domain all-weather road perception system based on vibration sensing, comprising:

[0048] A data acquisition module, used to acquire vibration sensing data and various roadside detection data, wherein the vibration sensing data and the roadside detection data both include time and space information;

[0049] A synchronization correction module, used to correct and synchronize the time and space information of the roadside detection data based on the time and space information of the vibration sensing data, so as to obtain the synchronous spatial information corresponding to the roadside detection data;

[0050] A data fusion module, used to fuse the vibration sensing data and the roadside detection data based on the spatiotemporal information of the vibration sensing data and the synchronization information data, and obtain a multi-dimensional information data set corresponding to each monitored vehicle;

[0051] The event perception module is used to perform full-scene perception based on the multi-dimensional information data set to obtain event perception results.

[0052] The present invention provides a method and system for all-weather road perception in all time and space domains based on vibration sensing, wherein the method obtains vibration sensing data that is less affected by bad weather and other roadside detection data that are easily affected by weather, corrects and synchronizes other roadside detection data based on the time and space information of the vibration sensing data, and fuses the various roadside detection data according to the synchronized spatial information obtained after synchronization, so as to achieve full-scene perception in all weather conditions. Compared with the prior art, the present invention performs synchronous calibration on various roadside detection data based on vibration sensing data, reduces the impact of bad weather on the accuracy of sensing data, and solves the problem in the prior art that the perception of the situation on the highway is seriously affected by weather. At the same time, the present invention also achieves accurate perception of the highway in all time and space domains by fusing vibration sensing data and roadside detection data, and utilizing the characteristics of two different detection methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 A method flow chart of an embodiment of a full-time and space-domain all-weather road perception method based on vibration sensing provided by the present invention;

[0054] Figure 2 for Figure 1 A method flow chart of an embodiment of step S102;

[0055] Figure 3 for Figure 1 A method flow chart of an embodiment of step S103;

[0056] Figure 4 for Figure 3 A method flow chart of an embodiment of step S302;

[0057] Figure 5 The present invention provides a system architecture diagram of an embodiment of a full-time and space-domain all-weather road perception system based on vibration sensing. DETAILED DESCRIPTION

[0058] The preferred embodiments of the present invention are described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not used to limit the scope of the present invention.

[0059] In the description of the present application, “plurality” means two or more than two, unless otherwise clearly and specifically defined.

[0060] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present invention. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0061] Because vibration sensor data is less affected by weather, this application synchronously processes other roadside detection data based on the spatiotemporal information of the vibration sensor data, so that other roadside detection data can eliminate the deviation caused by weather by themselves, and fuse all data for detection, thereby realizing the full spatiotemporal domain collection and management of highway traffic information, and greatly improving the ability to resist severe weather during perception.

[0062] The present invention provides a full-time and space-domain all-weather road perception method and system based on vibration sensing, which are described below respectively.

[0063] Combination Figure 1 As shown, a specific embodiment of the present invention discloses a vibration sensing-based all-time and all-weather road perception method, the method comprising:

[0064] S101, acquiring vibration sensing data and multiple roadside detection data, wherein the vibration sensing data and the roadside detection data both include time and space information;

[0065] S102, using the spatiotemporal information of the vibration sensing data as a reference, correcting and synchronizing the spatiotemporal information of the roadside detection data to obtain synchronized spatial information corresponding to the roadside detection data;

[0066] S103, based on the spatiotemporal information of the vibration sensing data and the synchronization information data, the vibration sensing data and the roadside detection data are integrated to obtain a multi-dimensional information data set corresponding to each monitored vehicle;

[0067] S104: Perform full-scene perception based on the multi-dimensional information data set to obtain event perception results.

[0068] The present invention provides a method and system for all-weather road perception in all time and space domains based on vibration sensing, wherein the method obtains vibration sensing data that is less affected by bad weather and other roadside detection data that is easily affected by weather, corrects and synchronizes other roadside detection data based on the spatiotemporal information of the vibration sensing data, and fuses the various roadside detection data based on the synchronized spatial information obtained after synchronization to achieve full-scene perception. Compared with the prior art, the present invention performs synchronous calibration on various roadside detection data based on vibration sensing data, reduces the impact of bad weather on the accuracy of sensing data, and solves the problem in the prior art that the perception of highway conditions is seriously affected by weather.

[0069] On the other hand, vibration sensors are mainly distributed (such as fiber grating array sensors), which can be laid throughout the highway for global collection, but vibration sensors cannot directly obtain individual information of vehicles, such as license plates, models, etc., and can only obtain vibration information of the road. The remaining roadside sensors are point-type (such as cameras, radars, speedometers, etc.), which can only be discretely set at a certain node on the highway and cannot achieve global collection, but they can accurately obtain individual information of the above vehicles, such as license plates, etc.

[0070] Therefore, in the method and system provided by the present invention, the two types of sensing methods, vibration sensing and other roadside sensing, are combined by means of fusing vibration sensing data and roadside detection data based on vibration sensing data. This allows the detection process to have the advantages of a complete acquisition domain and strong anti-interference ability of a vibration sensor, as well as the precision and flexibility of a roadside sensor, thereby achieving accurate perception of the entire time and space domain of the highway.

[0071] As a preferred embodiment, step S101 in this embodiment, obtaining vibration sensing data and various roadside detection data, both of which include spatiotemporal information, is mainly collected by roadside sensors, wherein the vibration sensing data is mainly obtained by fiber gratings buried throughout the highway (it can also be other array vibration sensing equipment), and other various roadside detection data are mainly obtained by deploying lidar, millimeter wave radar, video surveillance, license plate recognition, V2X-RSU, pre-weighing system, pre-transaction system, etc. at intervals of 2 km in toll plazas, bridges, long slopes, tunnels, service areas, ramp entrances and exits, and straight sections.

[0072] Among them, the spatial reference coordinate system of the vibration sensor data is a kilometer pile positioning system, and the spatial reference coordinate systems of the various roadside detection data include the absolute coordinate system and the relative coordinate system.

[0073] In a more specific embodiment of the present invention, the vibration sensing data obtained by a vibration sensing device similar to an array uses a kilometer pile positioning system, and the kilometer pile positioning coordinates under the kilometer pile positioning include the pile number and the lane number; the roadside detection data obtained by equipment such as a license plate recognition camera uses a relative coordinate system, and the relative coordinates under the relative coordinate system include the equipment installation position and the relative position coordinates; the roadside sensing data obtained by equipment such as radar uses an absolute coordinate system, that is, a geodetic coordinate system, and the absolute coordinates under the absolute coordinate system include longitude and latitude. Examples of specific data types are as follows:

[0074] At 16:12:30 on October 10, 2022, the data reported by each sensing device are as follows:

[0075] Vibration sensing data acquired by the array vibration sensing device:

[0076] Timestamp: 2022-10-10 16:12:30

[0077] Speed: 30 m / s

[0078] Lane Number: 3

[0079] Station number: K3+818

[0080] Vehicle type: 0 (car)

[0081] Roadside detection data obtained by the license plate recognition sensing device:

[0082] Installation location: K3+800

[0083] Timestamp: 2022-10-10 16:12:30

[0084] License plate number: Jaw A10001

[0085] License plate color: blue

[0086] Lane Number: 3

[0087] Horizontal distance (x): 18 meters

[0088] Vertical distance (y): 1.6 meters

[0089] Roadside detection data obtained by radar sensing equipment:

[0090] Timestamp: 2022-10-10 16:12:30

[0091] Speed: 30 m / s

[0092] Longitude: 114.9372004

[0093] Latitude: 30.29717446

[0094] Object Length: 4.6 meters

[0095] Object width: 1.8 meters

[0096] Subject height: 1.4 m

[0097] After obtaining the above data, it is necessary to perform synchronous correction processing on the above data to eliminate the data deviation caused by equipment such as radar and surveillance cameras that are easily affected by weather. Figure 2 As shown, as a preferred embodiment, step S102 in this embodiment, based on the spatiotemporal information of the vibration sensing data, correcting and synchronizing the spatiotemporal information of the roadside detection data to obtain the synchronous spatial information corresponding to the road test detection data, includes:

[0098] S201, performing time calibration on the time information in the vibration sensing data and the roadside detection data;

[0099] S202, establishing a correction mapping set according to the kilometer pile positioning system and the absolute coordinate system, wherein the correction mapping set includes a mapping relationship between reference kilometer pile positioning coordinates in the kilometer pile positioning system and reference absolute coordinates in the absolute coordinate system;

[0100] S203. Synchronize the relative coordinates of the roadside detection data based on the relative coordinate system according to the correction mapping set and obtain the synchronized spatial information, wherein the synchronized spatial information includes the reference kilometer pile positioning coordinates and the reference absolute coordinates.

[0101] In the above process, based on the correction mapping set, the spatial information of the roadside detection data is corrected and synchronized into the reference kilometer pile positioning coordinates and the reference absolute coordinates pre-set based on the vibration sensor data. This not only realizes the coordinate conversion under different coordinate systems, but also realizes the correction alignment of different spatial information, eliminating the deviation influence of the roadside detection data.

[0102] Specifically, in a preferred embodiment, step S202 in the above process, establishing a correction mapping set according to the kilometer pile positioning system and the absolute coordinate system, includes:

[0103] According to the kilometer pile positioning system, a plurality of the reference kilometer pile positioning coordinates are obtained, and the actual lane area corresponding to each of the reference kilometer pile positioning coordinates is obtained;

[0104] Selecting an actual position object from each of the actual lane areas;

[0105] Based on the absolute coordinate system, obtaining the absolute coordinates of each of the actual position objects;

[0106] The absolute coordinates of the actual position object corresponding to each of the reference kilometer pile positioning coordinates are used as the reference absolute coordinates to obtain the correction mapping set.

[0107] In a more specific embodiment, the lane centerline point corresponding to each reference kilometer pile positioning coordinate obtained by actual measurement is used as the actual position object, and the absolute coordinates of the lane centerline point are used as the reference absolute coordinates. Taking a single lane as an example, the correction mapping set established by the above steps is expressed as follows:

[0108] Set=Each(Ki*x0,(lat i ,lon i ))

[0109] Where Set is the correction mapping set, Ki is the reference kilometer pile point in the i-th reference kilometer pile positioning coordinates, X0 represents the distance between the kilometer pile points in the two reference kilometer pile positioning coordinates, and the constant, lat i is the longitude value of the reference absolute coordinate corresponding to the i-th reference kilometer pile positioning coordinate, lon i It is the dimension value of the reference absolute coordinate corresponding to the i-th reference kilometer pile positioning coordinate.

[0110] It is understandable that, in practice, the actual position object can also be selected based on other criteria, and for multi-lane data, it is only necessary to add the reference lane number in the corresponding reference kilometer pile positioning coordinates to the above-mentioned correction mapping set. Those skilled in the art can think of the implementation method based on the above description, and no further explanation is given in this article.

[0111] Further, as a preferred embodiment, step S203 in the above process, synchronizing the relative coordinates in the roadside detection data based on the relative coordinate system and obtaining the synchronized spatial information according to the correction mapping set, specifically includes:

[0112] According to the installation position of the equipment, the actual kilometer pile positioning coordinates corresponding to each relative coordinate in the kilometer pile positioning system are obtained;

[0113] According to the actual kilometer pile positioning coordinates, selecting the reference kilometer pile positioning coordinates having the smallest distance from the actual kilometer pile positioning coordinates from the correction mapping set;

[0114] According to the selected reference kilometer pile positioning coordinates and the corresponding reference absolute coordinates, the synchronous spatial information corresponding to the relative coordinates is obtained.

[0115] A specific formula for the above process is as follows:

[0116] i=Min(|KK i |)

[0117] lat,lon=Set(K i )

[0118] Where Ki is the reference kilometer pile point in the selected i-th reference kilometer pile positioning coordinates, K is the actual kilometer pile positioning coordinates corresponding to the target relative coordinates currently being synchronized, lat and lon are the longitude and latitude values ​​in the reference absolute coordinates corresponding to the selected reference kilometer pile positioning coordinates, that is, the synchronization space information corresponding to the target relative coordinates.

[0119] The above step S203 is described below through a more specific embodiment:

[0120] Based on the specific data described above, the actual positioning point, i.e., the actual kilometer pile positioning coordinates, is first calculated based on the horizontal (x) and vertical (y) coordinates relative to the installation position reported by the license plate recognition sensing device. In this embodiment, the actual kilometer pile positioning coordinates of the object sensed by the license plate recognition camera according to the conversion method are: K3+800+18=K3+818, and the lane number is 3;

[0121] In the correction mapping set, the spacing of the benchmark kilometer posts is one meter. The nearest mapping point to K3+818 is K3+818 when querying the coordinate mapping set. The corresponding longitude and latitude values ​​are 114.9372004 longitude and 30.29717446 latitude. The above K3+818, longitude 114.9372004, and latitude 30.29717446 are the synchronous spatial information corresponding to the roadside detection data obtained by the license plate recognition sensing device.

[0122] Further, combined with Figure 3 As shown, as a preferred embodiment, step S103 in this embodiment, based on the spatiotemporal information of the vibration sensor data and the synchronization information data, fuses the vibration sensor data and the roadside detection data, and obtains a multidimensional information data set corresponding to each monitored vehicle, specifically including:

[0123] S301, establishing a multi-dimensional information data model according to the vibration sensing data and the drive test detection data;

[0124] S302, matching the time information of the roadside detection data and the corresponding synchronous spatial information with the time information and spatial information of the vibration sensor data respectively, and fusing the successfully matched roadside detection data and the vibration sensor data based on the multi-dimensional information data model to obtain the multi-dimensional information data set corresponding to each monitored vehicle;

[0125] S303, assigning a unique identification ID to the corresponding monitored vehicle according to the monitored vehicle corresponding to the multi-dimensional information data set;

[0126] S304: Track the data of the monitored vehicle and update the corresponding multi-dimensional information data set according to the unique identification ID.

[0127] As a preferred embodiment, the roadside detection data in this embodiment includes relative detection data based on the relative coordinate system (i.e., data obtained by the sensing device mentioned above) and absolute detection data based on the absolute coordinate system (i.e., data obtained by the radar sensing device mentioned above). Figure 4 As shown, step S302 in the above process, matching the time information of the roadside detection data and the corresponding synchronous spatial information with the time information and spatial information of the vibration sensor data respectively, and fusing the successfully matched roadside detection data and the vibration sensor data based on the multi-dimensional information data model to obtain the multi-dimensional information data set corresponding to each monitored vehicle, specifically includes:

[0128] S401, matching the time information of the relative detection data and the corresponding synchronous spatial information with the time information and spatial information of the vibration sensor data respectively, and fusing the successfully matched relative detection data and the vibration sensor data based on the multi-dimensional information data model to obtain an initial multi-dimensional information data set corresponding to each monitored vehicle;

[0129] S402. Match the time information and space information of the absolute detection data with the time information and space information of the initial multidimensional information data set respectively, and based on the multidimensional information data model, fuse the successfully matched absolute detection data and the initial alignment information data set to obtain the multidimensional information data set corresponding to each monitored vehicle.

[0130] In a preferred embodiment, the above step S401, respectively matching the time information of the relative detection data and the corresponding synchronous spatial information with the time information and spatial information of the vibration sensor data, and fusing the successfully matched relative detection data and the vibration sensor data based on the multi-dimensional information data model to obtain an initial multi-dimensional information data set corresponding to each monitored vehicle, specifically includes:

[0131] Matching the time information of the relative detection data with the time information of the vibration sensing data to obtain time-matched relative data;

[0132] Matching the synchronous spatial information corresponding to the time matching relative data with the spatial information of the vibration sensing data to obtain time-space matching relative data;

[0133] Based on the multi-dimensional information data model, the time-space matching relative data and the vibration sensing data are integrated to obtain the initial multi-dimensional information data set;

[0134] The time information of the initial multi-dimensional information data set is the time information of the relative detection data, and the space information of the initial multi-dimensional information data set is the matching synchronous space information.

[0135] Further, in a preferred embodiment, the above step S402, matching the time information and space information of the absolute detection data with the time information and space information of the initial multidimensional information data set respectively, and fusing the successfully matched absolute detection data and the initial alignment information data set based on the multidimensional information data model to obtain the multidimensional information data set corresponding to each monitored vehicle, specifically includes:

[0136] Matching the time information of the absolute detection data with the time information of the initial multi-dimensional information data set to obtain time-matched absolute data;

[0137] Matching the spatial information of the time-matched absolute data with the spatial information of the initial multi-dimensional information data set to obtain time-space matched absolute data;

[0138] Based on the multidimensional information data model, the spatiotemporal matching absolute data and the initial multidimensional information data set are fused to obtain the multidimensional information data set.

[0139] The present invention also provides a more specific embodiment to illustrate the above steps S301 to S304:

[0140] The multidimensional information model of the perception object constructed in this embodiment is as follows:

[0141]

[0142]

[0143] In this embodiment, the timestamp 2022-10-10 16:12:30 (i.e., the time information of the vibration sensor data) in the vibration sensor data is taken, and the data matching the time dimension is searched in the license plate recognition camera data. In this example, a record is matched (i.e., time matching relative data); then the object position in the array sensor data (i.e., the spatial information of the vibration sensor data) is obtained, which is the pile number K3+818 and the lane number 3 in this example, and the position in the synchronous spatial information corresponding to the matched license plate recognition camera data is compared to see whether it is consistent. In this example, it is completely consistent, so the vehicle objects perceived by these two records can be marked as the same perceived object. At this time, the matched license plate camera data and its corresponding synchronous spatial information can be used as the time-space matching relative data, and the two perception source data are fused. The vehicle position, speed, and size are obtained from the data reported by the array vibration sensing facility, and the vehicle license plate number and license plate color are obtained from the data reported by the license plate recognition camera. In addition, a unique object ID is assigned to the perceived object, and a piece of data in the fused initial multi-dimensional information data set is:

[0144] Object ID 1 Timestamp 2022-10-10 16:12:30 License plate number Jaw A10001 License plate color blue speed 108Km / h Lane Number 3 Pile No. K3+818 longitude 114.9372004 latitude 30.29717446 length width high

[0145] The next step is to further fuse the radar data (i.e., absolute detection data) based on the initial multi-dimensional information data after fusion, take the timestamp 2022-10-10 16:12:30 in the fused data, and find out whether there is a matching record in the radar perception data. In this case, there is a matching record (i.e., time matching absolute data); then perform position matching of the perceived object. The radar uses the positioning coordinate system as the absolute coordinate system, so the converted longitude and latitude data in the fused data (spatial information of the multi-dimensional information data set) is compared with the matching record of the radar perception data. In this case, they are completely consistent, so it can be considered that the radar perception object and the object corresponding to the initial multi-dimensional information data set are the same perception object. At this time, the matched radar perception data can be used as the absolute data for time-space matching, and the radar perception data can be fused. The multi-dimensional information data obtained after fusion is:

[0146]

[0147]

[0148] According to the above process, based on a unified object, after the data of each timestamp are integrated, the multi-dimensional information data set corresponding to the vehicle can be obtained.

[0149] Then, based on the object ID, the bound target vehicle to be monitored can be tracked in the entire area, and a unique identification ID can be set for the perception object that has been fused. In the subsequent perception data processing, the object ID will no longer change and the entire road section will be continuously tracked. The newly acquired data will continue to be integrated into the existing multi-dimensional information data set. The specific tracking and fusion process also uses grating fiber sensors for full-area tracking and corrections at nodes where other sensors are deployed. It is the same as the method described above, so this article will not go into details.

[0150] After obtaining the multi-dimensional information data set corresponding to each vehicle, step S104 can be performed. As a preferred embodiment, step S104 in this embodiment, performing full scene perception according to the multi-dimensional information data set to obtain event perception results, specifically includes:

[0151] Construct a set of vehicle behavior patterns and obtain mathematical characteristics of each vehicle behavior pattern;

[0152] Obtaining a vehicle behavior mathematical representation model according to the vehicle behavior pattern set and the mathematical characteristics;

[0153] According to the multi-dimensional information data set, based on the vehicle behavior mathematical characterization model, the event perception result is obtained.

[0154] A more detailed embodiment is given below to illustrate the above step S104:

[0155] According to the characteristics of road vehicle operation, typical vehicle behavior patterns are proposed, such as acceleration, deceleration, lane change, etc., and a collection of typical vehicle behavior patterns is constructed;

[0156] Select specific parameters, such as vehicle speed, acceleration, etc., to define the mathematical characteristics of the vehicle's behavior pattern;

[0157] Combined with available roadside detection data, a mathematical representation model of vehicle behavior and information is established;

[0158] Using alternative safety indicators such as TTC, based on traffic conflict theory and technology, the vehicle traffic risk is quantitatively described and graded scientifically and demonstrably, and the risk level is calculated using the target vehicle and its surrounding information;

[0159] According to the parameter requirements of behavior pattern recognition and risk level calculation, multi-source fusion information of the target vehicle is collected and extracted, and the established representation relationship is used to complete the vehicle behavior pattern recognition.

[0160] Integrate regional identification and calculation results to achieve behavioral pattern recognition and dangerous state identification of vehicles on the road network (i.e. event perception results).

[0161] For example, in this embodiment, the acquired multidimensional information data set is as follows:

[0162]

[0163]

[0164] The speeding behavior pattern is established based on the vehicle speed. The speed is the behavior characteristic of the pattern. Whether the vehicle speed exceeds 120Km / h is judged as a mathematical representation model. If it exceeds, the speeding state is recorded. If the speeding state of the same object has been recorded in the speeding state record list, the vehicle speeding event is reported. In the above example, the object speed does not exceed 120Km / h, so there is no need to record the speeding state.

[0165] Similarly, to establish the behavior pattern of illegal parking, the mathematical characteristic of illegal parking is also speed, and its status judgment is also based on the vehicle speed data in the multidimensional information data set. In the above example, the vehicle speed of the third record is 0, and it is recorded as illegal parking at this time. At the same time, the illegal parking record list is queried to see if there is an illegal parking status record for the same object. If it is confirmed that there is no illegal parking status record for the same object in the illegal parking record list, the illegal parking status record of the object is added to the list; for the 9th record, the vehicle speed is still 0. Since there is an illegal parking status record in the record list, the time stamps of the two record data are compared. If the time difference is greater than 5 seconds, the illegal parking event is reported.

[0166] The continuous lane change status is determined based on the lane number data in the multidimensional information data set. In the above example, starting from the first record, after the data fusion is completed, the continuous lane change status record list is queried to see if there is a record of the same object. If no record is found, the object is added to the continuous lane change status record list; starting from the second record, if there is an existing record, the lane numbers of the two records are compared to see if they are the same. If they are the same, the timestamp in the record list is updated to the timestamp of the latest data; when it is the 11th record, the lane number is changed to 2, and the first lane change is recorded; when it is the 14th record, the lane number is changed to 3 again, and the first lane change record has been queried in the record list. The time difference between the two lane changes is calculated, and if it does not exceed 3 seconds, the vehicle continuous lane change event is reported.

[0167] After obtaining the above event perception results, you can perform the following processing:

[0168] Release vehicle driving information and emergencies to drivers, passengers, highway police, and road administration;

[0169] Statistics of road network information and environmental information, for example, analyzing the number of vehicles on the road, traffic density, predicting future traffic volume, using gantry data to analyze road network traffic volume characteristics, digging out hidden factor relationships and development trends from toll data, using factor analysis to classify all factors that may affect traffic flow into several groups for evaluating factors affecting traffic flow; using causal analysis to select several key risk factors that have a greater impact on the final result; designing a highway traffic flow analysis and prediction model, using recursive algorithm iterative training to continuously improve the traffic flow prediction model, improve the accuracy of traffic flow prediction, and push congestion prevention and smooth traffic to advance preparation. And publish the operation status information such as congestion events and expected congestion time to highway police, road administration, and drivers and passengers through information boards and APPs.

[0170] From the above description, it can be seen that the present invention intends to adopt the grating fiber sensing technology with relatively low cost, lane-level information collection, and no influence from climatic conditions. By integrating high-definition video detectors, radar and grating fiber sensing technology, it realizes the full-time and space-domain collection and management of highway traffic information, and achieves the following breakthroughs:

[0171] 1. Realize full-area vehicle tracking, making up for the blind spots of traditional monitoring methods;

[0172] 2. Real-time monitoring of vehicle trajectory and driving speed, timely warning, and providing evidence for accidents;

[0173] 3. Independent of lighting conditions, monitoring road network information around the clock and in all weather conditions;

[0174] 4. Discover accidents promptly and locate them accurately, changing passive alarm reception to active alarm response.

[0175] Most importantly, the present invention provides a method and system for all-weather road perception in all time and space domains based on vibration sensing, wherein the method obtains vibration sensing data that is less affected by bad weather and other roadside detection data that is easily affected by weather, corrects and synchronizes the roadside detection data based on the spatiotemporal information of the vibration sensing data, and fuses the multiple roadside detection data based on the synchronized spatial information obtained after synchronization to achieve full-scene perception. Compared with the prior art, the present invention synchronizes and calibrates multiple roadside detection data based on vibration sensing data, eliminates the impact of bad weather on the accuracy of sensing data, and solves the problem in the prior art that the perception of highway conditions is seriously affected by weather.

[0176] In order to better implement the all-time and space-domain all-weather road perception method based on vibration sensing in the embodiment of the present invention, on the basis of the all-time and space-domain all-weather road perception method based on vibration sensing, correspondingly, please refer to Figure 5 , Figure 5This is a schematic diagram of the structure of an embodiment of a full-time and space-domain all-weather road perception system based on vibration sensing provided by the present invention. An embodiment of the present invention provides a full-time and space-domain all-weather road perception system 500 based on vibration sensing, including:

[0177] A data acquisition module 510, for acquiring vibration sensing data and a variety of roadside detection data, wherein the vibration sensing data and the roadside detection data both include time and space information;

[0178] A synchronization correction module 520, used to correct and synchronize the time and space information of the roadside detection data based on the time and space information of the vibration sensing data, and obtain synchronized spatial information corresponding to the roadside detection data;

[0179] A data fusion module 530 is used to fuse the vibration sensing data and the roadside detection data based on the spatiotemporal information of the vibration sensing data and the synchronization information data, and obtain a multi-dimensional information data set corresponding to each monitored vehicle;

[0180] The event perception module 540 is used to perform full-scene perception based on the multi-dimensional information data set to obtain event perception results.

[0181] It should be noted here that the corresponding system provided in the above embodiments can implement the technical solutions described in the above method embodiments. The specific implementation principles of the above modules or units can be found in the corresponding contents in the above method embodiments, which will not be repeated here.

[0182] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. A vibration sensing-based all-time and all-weather road perception method, characterized in that: include: Acquiring vibration sensing data and a plurality of roadside detection data, wherein the vibration sensing data and the roadside detection data both include temporal and spatial information; Based on the spatiotemporal information of the vibration sensing data, the spatiotemporal information of the roadside detection data is corrected and synchronized to obtain synchronized spatial information corresponding to the roadside detection data; Based on the spatiotemporal information and synchronous information data of the vibration sensing data, the vibration sensing data and the roadside detection data are integrated to obtain a multi-dimensional information data set corresponding to each monitored vehicle; Perform full scene perception based on the multi-dimensional information data set to obtain event perception results; The spatial reference coordinate system of the vibration sensor data is a kilometer pile positioning system, and the spatial reference coordinate systems of the multiple roadside detection data include an absolute coordinate system and a relative coordinate system; the time and space information of the roadside detection data is corrected and synchronized based on the time and space information of the vibration sensor data to obtain the synchronized spatial information corresponding to the roadside detection data, including: Performing time calibration on the time information in the vibration sensing data and the roadside detection data; According to the kilometer pile positioning system and the absolute coordinate system, a correction mapping set is established, wherein the correction mapping set includes a mapping relationship between reference kilometer pile positioning coordinates under the kilometer pile positioning system and reference absolute coordinates under the absolute coordinate system; According to the correction mapping set, synchronizing the relative coordinates based on the relative coordinate system in the roadside detection data and obtaining the synchronized spatial information, wherein the synchronized spatial information includes the reference kilometer pile positioning coordinates and the reference absolute coordinates; The kilometer pile positioning coordinates include a pile number and a lane number, and the absolute coordinates of the absolute coordinate system include longitude and latitude; and establishing a correction mapping set according to the kilometer pile positioning system and the absolute coordinate system includes: According to the kilometer pile positioning system, a plurality of the reference kilometer pile positioning coordinates are obtained, and the actual lane area corresponding to each of the reference kilometer pile positioning coordinates is obtained; Selecting an actual position object from each of the actual lane areas; Based on the absolute coordinate system, obtaining the absolute coordinates of each of the actual position objects; The absolute coordinates of the actual position object corresponding to each of the reference kilometer pile positioning coordinates are used as the reference absolute coordinates to obtain the correction mapping set.

2. The vibration sensing-based all-time and all-weather road perception method according to claim 1, characterized in that: The spatial information based on the relative coordinate system in the roadside detection data includes relative coordinates in the relative coordinate system, and the relative coordinates include equipment installation positions and relative position coordinates; and synchronizing the relative coordinates based on the relative coordinate system in the roadside detection data according to the correction mapping set and obtaining the synchronized spatial information includes: According to the installation position of the equipment, the actual kilometer pile positioning coordinates corresponding to each relative coordinate in the kilometer pile positioning system are obtained; According to the actual kilometer pile positioning coordinates, selecting the reference kilometer pile positioning coordinates having the smallest distance from the actual kilometer pile positioning coordinates from the correction mapping set; According to the selected reference kilometer pile positioning coordinates and the corresponding reference absolute coordinates, the synchronous spatial information corresponding to the relative coordinates is obtained.

3. The vibration sensing-based all-time and all-weather road perception method according to claim 1, characterized in that: The method of fusing the vibration sensing data and the roadside detection data based on the spatiotemporal information of the vibration sensing data and the synchronization information data to obtain a multi-dimensional information data set corresponding to each monitored vehicle includes: Establishing a multi-dimensional information data model based on the vibration sensing data and the drive test detection data; Matching the time information of the roadside detection data and the corresponding synchronous spatial information with the time information and spatial information of the vibration sensor data respectively, and fusing the successfully matched roadside detection data and the vibration sensor data based on the multi-dimensional information data model to obtain the multi-dimensional information data set corresponding to each monitored vehicle; According to the monitored vehicle corresponding to the multi-dimensional information data set, a unique identification ID is assigned to the corresponding monitored vehicle; According to the unique identification ID, the data of the monitored vehicle is tracked and the corresponding multi-dimensional information data set is updated.

4. The vibration sensing-based all-time and all-weather road perception method according to claim 3 is characterized in that: The roadside detection data includes relative detection data based on the relative coordinate system and absolute detection data based on the absolute coordinate system; the time information of the roadside detection data and the corresponding synchronous spatial information are matched with the time information and spatial information of the vibration sensor data respectively, and based on the multi-dimensional information data model, the successfully matched roadside detection data and the vibration sensor data are integrated to obtain the multi-dimensional information data set corresponding to each monitored vehicle, including: Matching the time information of the relative detection data and the corresponding synchronous spatial information with the time information and spatial information of the vibration sensor data respectively, and fusing the successfully matched relative detection data and the vibration sensor data based on the multi-dimensional information data model to obtain an initial multi-dimensional information data set corresponding to each monitored vehicle; The time information and space information of the absolute detection data are matched with the time information and space information of the initial multidimensional information data set respectively, and based on the multidimensional information data model, the successfully matched absolute detection data and the initial multidimensional information data set are fused to obtain the multidimensional information data set corresponding to each monitored vehicle.

5. The vibration sensing-based all-time and all-weather road perception method according to claim 4, characterized in that: The time information of the relative detection data and the corresponding synchronous spatial information are matched with the time information and spatial information of the vibration sensor data respectively, and based on the multi-dimensional information data model, the successfully matched relative detection data and the vibration sensor data are integrated to obtain an initial multi-dimensional information data set corresponding to each monitored vehicle, including: Matching the time information of the relative detection data with the time information of the vibration sensing data to obtain time-matched relative data; Matching the synchronous spatial information corresponding to the time matching relative data with the spatial information of the vibration sensing data to obtain time-space matching relative data; Based on the multidimensional information data model, the time-space matching relative data and the vibration sensor data are integrated to obtain the initial multidimensional information data set; The time information of the initial multi-dimensional information data set is the time information of the relative detection data, and the space information of the initial multi-dimensional information data set is the matching synchronous space information.

6. The vibration sensing-based all-time and all-weather road perception method according to claim 5, characterized in that: The time information and space information of the absolute detection data are matched with the time information and space information of the initial multi-dimensional information data set respectively, and based on the multi-dimensional information data model, the successfully matched absolute detection data and the initial alignment information data set are fused to obtain the multi-dimensional information data set corresponding to each monitored vehicle, including: Matching the time information of the absolute detection data with the time information of the initial multi-dimensional information data set to obtain time-matched absolute data; Matching the spatial information of the time-matched absolute data with the spatial information of the initial multi-dimensional information data set to obtain time-space matched absolute data; Based on the multidimensional information data model, the spatiotemporal matching absolute data and the initial multidimensional information data set are fused to obtain the multidimensional information data set.

7. The vibration sensing-based all-time and all-weather road perception method according to claim 1, characterized in that: The performing of full scene perception according to the multi-dimensional information data set to obtain an event perception result includes: Construct a set of vehicle behavior patterns and obtain mathematical characteristics of each vehicle behavior pattern; Obtaining a vehicle behavior mathematical representation model according to the vehicle behavior pattern set and the mathematical characteristics; According to the multi-dimensional information data set, based on the vehicle behavior mathematical characterization model, the event perception result is obtained.

8. A full-time and all-weather road perception system based on vibration sensing, characterized in that: include: A data acquisition module, used to acquire vibration sensing data and various roadside detection data, wherein the vibration sensing data and the roadside detection data both include time and space information; A synchronization correction module, used to correct and synchronize the time and space information of the roadside detection data based on the time and space information of the vibration sensing data, so as to obtain synchronized spatial information corresponding to the roadside detection data; A data fusion module, used to fuse the vibration sensing data and the roadside detection data based on the spatiotemporal information synchronization information data of the vibration sensing data, and obtain a multi-dimensional information data set corresponding to each monitored vehicle; An event perception module, used to perform full-scene perception based on the multi-dimensional information data set to obtain event perception results; The spatial reference coordinate system of the vibration sensor data is a kilometer pile positioning system, and the spatial reference coordinate systems of the multiple roadside detection data include an absolute coordinate system and a relative coordinate system; the time and space information of the roadside detection data is corrected and synchronized based on the time and space information of the vibration sensor data to obtain the synchronized spatial information corresponding to the roadside detection data, including: Performing time calibration on the time information in the vibration sensing data and the roadside detection data; According to the kilometer pile positioning system and the absolute coordinate system, a correction mapping set is established, wherein the correction mapping set includes a mapping relationship between reference kilometer pile positioning coordinates under the kilometer pile positioning system and reference absolute coordinates under the absolute coordinate system; According to the correction mapping set, synchronizing the relative coordinates based on the relative coordinate system in the roadside detection data and obtaining the synchronized spatial information, wherein the synchronized spatial information includes the reference kilometer pile positioning coordinates and the reference absolute coordinates; The kilometer pile positioning coordinates include a pile number and a lane number, and the absolute coordinates of the absolute coordinate system include longitude and latitude; and establishing a correction mapping set according to the kilometer pile positioning system and the absolute coordinate system includes: According to the kilometer pile positioning system, a plurality of the reference kilometer pile positioning coordinates are obtained, and the actual lane area corresponding to each of the reference kilometer pile positioning coordinates is obtained; Selecting an actual position object from each of the actual lane areas; Based on the absolute coordinate system, obtaining the absolute coordinates of each of the actual position objects; The absolute coordinates of the actual position object corresponding to each of the reference kilometer pile positioning coordinates are used as the reference absolute coordinates to obtain the correction mapping set.

Images