CN-116244624-B - Mobile object track extraction method and device, electronic equipment and medium

CN116244624BCN 116244624 BCN116244624 BCN 116244624BCN-116244624-B

Abstract

The invention relates to a method, a device, electronic equipment and a medium for extracting a moving object track, which comprise the steps of acquiring multidimensional space-time signal data based on a grating sensing array, preprocessing the multidimensional space-time signal data to obtain a multidimensional space-time signal matrix, analyzing a connected domain of the multidimensional space-time signal matrix to obtain a plurality of different connected domains, linearly fitting the track of the moving object in each connected domain of the plurality of different connected domains, and matching the track position of the moving object with the data of a previous frame to obtain a plurality of complete tracks. The invention extracts the track of the moving object by using the grating array sensing technology, and can output the position and the speed of the moving object in real time.

Inventors

TONG MENG
XIONG XIN
MA JUNJIE
SONG KE
XU YIMIN
WANG YUEMING

Assignees

武汉烽理光电技术有限公司
武汉理工大学

Dates

Publication Date: 20260512
Application Date: 20221230

Claims (7)

1. A moving object trajectory extraction method, comprising: Acquiring multidimensional space-time signal data based on a grating sensing array, and preprocessing to acquire a multidimensional space-time signal matrix; Carrying out connected domain analysis on the multidimensional space-time signal matrix to obtain a plurality of different connected domains; linearly fitting the trajectories of the moving objects in each of the plurality of different connected domains; matching the track position of the moving object with the data of the previous frame to obtain a plurality of complete tracks; The linear fitting of the trajectories of the moving objects in each of the plurality of different connected domains includes: Performing linear fitting on the track of the moving object in each of the different connected domains based on a Hough transform algorithm and a clustering algorithm; The linear fitting of the trajectories of the moving objects in each of the plurality of different connected domains based on the hough transform algorithm and the clustering algorithm includes: Taking out polar coordinates of the moving object in each of the different connected domains, the occurrence times of the track points of the moving object exceeding a preset threshold value, based on a Hough transform algorithm, and forming a first set; taking out the polar coordinates of the first set meeting the preset condition based on a clustering algorithm to obtain a second set; Converting all polar coordinates in the second set into straight lines under a rectangular coordinate system to obtain positions and average speeds of different moving objects within preset time, and repeating the above process for each connected domain, so that the positions and average speeds of all the moving objects in the whole domain are obtained in real time; Taking out polar coordinates of which the occurrence times of the track points of the moving object in each of the plurality of different connected domains exceeds a preset threshold value based on a Hough transformation algorithm to form a first set, wherein the polar coordinates comprise: converting the rectangular coordinates of the points in each of the plurality of different connected domains into polar coordinates; And taking out polar coordinates with the times of crossing the curve at a point being larger than a preset threshold value to form a first set.
2. The method of claim 1, wherein the acquiring multi-dimensional spatiotemporal signal data comprises: grating sensing array optical fibers are respectively distributed on each lane of the highway in the longitudinal direction extending along the lane, and each grating sensing array optical fiber comprises a plurality of fiber grating sensors; And acquiring continuous vibration signals of the moving object based on the plurality of fiber bragg grating sensors, wherein the continuous vibration signals form multidimensional space-time signal data.
3. The method for extracting a moving object trajectory according to claim 1, wherein the extracting, based on a clustering algorithm, the polar coordinates of the first set satisfying a preset condition to obtain a second set includes: the first step, a plurality of points are randomly selected based on the first set to obtain a plurality of initial clustering centers; dividing the points into clusters formed by the nearest initial cluster centers according to the distances from each point in the first set to the initial cluster centers; Thirdly, calculating the central point of each cluster, judging whether the central point is equal to the original initial cluster central point, if so, obtaining a first cluster central point, and if not, executing a second step; fourth, a second set formed by a plurality of first clustering center points is obtained.
4. The method for extracting a track of a moving object according to claim 1, wherein the matching the track position of the moving object with the data of the previous frame to obtain a plurality of complete tracks includes: Matching the track position of the moving object with the data of the previous frame; and obtaining the channel where the moving object is located according to the data distribution acquired by different channels at different moments, and determining a plurality of complete tracks.
5. A moving object trajectory extraction device, characterized by comprising: the multi-dimensional space-time signal matrix acquisition unit is used for acquiring multi-dimensional space-time signal data based on the grating sensing array and preprocessing the multi-dimensional space-time signal data to acquire a multi-dimensional space-time signal matrix; The connected domain acquisition unit is used for carrying out connected domain analysis on the multidimensional space-time signal matrix to obtain a plurality of different connected domains; A track fitting unit, configured to linearly fit the tracks of the moving objects in each of the plurality of different connected domains; the track acquisition unit is used for matching the track position of the moving object with the data of the previous frame to obtain a plurality of complete tracks; The linear fitting of the trajectories of the moving objects in each of the plurality of different connected domains includes: Performing linear fitting on the track of the moving object in each of the different connected domains based on a Hough transform algorithm and a clustering algorithm; The linear fitting of the trajectories of the moving objects in each of the plurality of different connected domains based on the hough transform algorithm and the clustering algorithm includes: Taking out polar coordinates of the moving object in each of the different connected domains, the occurrence times of the track points of the moving object exceeding a preset threshold value, based on a Hough transform algorithm, and forming a first set; taking out the polar coordinates of the first set meeting the preset condition based on a clustering algorithm to obtain a second set; Converting all polar coordinates in the second set into straight lines under a rectangular coordinate system to obtain positions and average speeds of different moving objects within preset time, and repeating the above process for each connected domain, so that the positions and average speeds of all the moving objects in the whole domain are obtained in real time; Taking out polar coordinates of which the occurrence times of the track points of the moving object in each of the plurality of different connected domains exceeds a preset threshold value based on a Hough transformation algorithm to form a first set, wherein the polar coordinates comprise: converting the rectangular coordinates of the points in each of the plurality of different connected domains into polar coordinates; And taking out polar coordinates with the times of crossing the curve at a point being larger than a preset threshold value to form a first set.
6. An electronic device comprising a memory and a processor, wherein, The memory is used for storing programs; the processor is coupled to the memory for executing the program stored in the memory to implement the steps of a moving object trajectory extraction method as claimed in any one of the preceding claims 1 to 4.
7. A computer readable storage medium storing a computer readable program or instructions which when executed by a processor is capable of carrying out the steps of a moving object trajectory extraction method according to any one of the preceding claims 1 to 4.

Description

Mobile object track extraction method and device, electronic equipment and medium Technical Field The present invention relates to the field of optical fiber sensing technologies, and in particular, to a method and apparatus for extracting a track of a moving object, an electronic device, and a medium. Background Along with the rapid improvement of productivity level, the positioning equipment, the mobile sensing equipment and the wireless communication equipment are continuously developed and widely applied, so that the way of acquiring various track data is greatly widened, the space-time track data of various moving objects can be conveniently and reliably acquired in real time to form signals, and the method has great significance in carrying out deep analysis on the signals and extracting valuable information. The moving target track identification and extraction is a typical multidisciplinary cross research problem related to traffic engineering and intelligent science and technology, and has important theoretical research and practical application values in the fields of intelligent traffic supervision, abnormal behavior detection, unmanned aircraft autonomous navigation and the like. Compared with the traditional sensing technology, the grating array sensing technology is a novel sensing technology integrating the advantages of distributed optical fibers and fiber gratings. The method has a plurality of natural advantages in traffic monitoring, and mainly comprises strong electromagnetic interference resistance. The optical fiber is anti-electromagnetic interference, stable in signal and small in interference in long-distance transmission, and is suitable for severe environments. The grating array sensing technology detects vibration signals distributed along the optical fiber, the vibration signals cannot be affected by light rays and extreme weather, and severe weather conditions such as night on a highway, heavy rain, snow, fog and the like cannot affect the tracking measurement of the vibration sensing on the moving object on the real-time track. In the traffic field, the running track of the vehicle can basically provide all dynamic information required by traffic, so that the extraction of the vehicle track is the primary target for researching most traffic problems, at present, the vehicle track identification and extraction are mainly realized through traffic monitoring videos or radars, the monitoring videos relate to the field of computer vision, the radars relate to the field of point cloud processing, but the monitoring videos and the radars have limited detection distances, the cost of global coverage is too high, the later processing is complex, and the difficulty of achieving real-time accurate vehicle detection and tracking is higher. Disclosure of Invention In view of the foregoing, it is necessary to provide a method, an apparatus, an electronic device and a medium for extracting a track of a moving object, so as to solve the problem in the prior art that it is difficult to accurately detect the position and the speed of the moving object in real time. In order to solve the above problems, the present invention provides a moving object trajectory extraction method, including: Acquiring multidimensional space-time signal data based on a grating sensing array, and preprocessing to acquire a multidimensional space-time signal matrix; Carrying out connected domain analysis on the multidimensional space-time signal matrix to obtain a plurality of different connected domains; linearly fitting the trajectories of the moving objects in each of the plurality of different connected domains; And matching the track position of the moving object with the data of the previous frame to obtain a plurality of complete tracks. In some possible implementations, the acquiring multi-dimensional spatio-temporal signal data includes: grating sensing array optical fibers are respectively distributed on each lane of the highway in the longitudinal direction extending along the lane, and each grating sensing array optical fiber comprises a plurality of fiber grating sensors; And acquiring continuous vibration signals of the moving object based on the plurality of fiber bragg grating sensors, wherein the continuous vibration signals form multidimensional space-time signal data. In some possible implementations, the linearly fitting the trajectories of the moving objects in each of the number of different connected domains includes: And linearly fitting the trajectories of the moving objects in each of the plurality of different connected domains based on a Hough transform algorithm and a clustering algorithm. In some possible implementations, the linear fitting of the trajectories of the moving objects in each of the several different connected domains based on the hough transform algorithm and the clustering algorithm includes: Taking out polar coordinates of the moving object in each of the different conne