Track segmentation method and system based on key point identification

Abstract

The invention discloses a trajectory segmentation method and system based on key point recognition, and relates to the field of trajectory analysis and pattern recognition, and the method comprises the steps: S1, obtaining multi-source trajectory data, and generating time sequence trajectory information; s2, calculating the acceleration corresponding to the track point according to the track information, further calculating the corresponding curvature, carrying out curvature related information statistics through the curvature of the track, and calculating an optimized curvature first-order derivative corresponding to the track point through the curvature information and the statistical information of the curvature; s3, marking acceleration key points, optimizing curvature first-order derivative key points and circling area key points; and S4, integrating information of the acceleration key point, the optimized curvature first-order derivative key point and the circling area key point, and generating a track segmentation result. According to the method, the key points in the trajectory can be accurately recognized, particularly, high performance can be kept in complex trajectories and high-noise data, and the trajectory analysis efficiency and reliability are effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of trajectory analysis and pattern recognition, and in particular to a trajectory segmentation method and system based on key point recognition. Background Art

[0002] The trajectory segmentation algorithm is a key tool for automatically analyzing multi-source trajectory data, aiming to segment complex and broken trajectories to automatically identify and segment significant change parts in the trajectories, obtain target behavior characteristics, and make the trajectory data present more meaningful segments. It is the basis for trajectory maneuver pattern recognition and analysis. The trajectory segmentation algorithm supports a large number of different types of trajectory data and meets the requirements of high-frequency motion analysis. In modern trajectory analysis and pattern recognition, accurately identifying key points in the trajectory and segmenting the trajectory are important links for realizing applications such as intelligent traffic management, unmanned aerial vehicle navigation, and motion analysis. However, in the prior art, most trajectory segmentation methods only rely on simple speed and position changes, lack a comprehensive analysis of the complex characteristics of the trajectory, and cannot accurately identify key points in the trajectory, resulting in low recognition accuracy under high-noise data and complex trajectory conditions.

[0003] Therefore, there is an urgent need to provide a solution for trajectory segmentation based on key point recognition. Summary of the Invention

[0004] To solve the above problems, the technical solution of the present invention provides a trajectory segmentation method and system based on key point recognition, which can improve the recognition accuracy of key points.

[0005] According to the first aspect embodiment of the technical solution of the present invention, a trajectory segmentation method based on key point recognition is provided, including: S1. Data preprocessing, obtaining multi-source trajectory data and generating time-series trajectory information; S2. Feature calculation, calculating the acceleration corresponding to the trajectory point according to the trajectory information, calculating the curvature corresponding to the trajectory point according to the trajectory information and the acceleration, performing curvature-related information statistics through the curvature of the trajectory, and calculating the optimized curvature first derivative corresponding to the trajectory point through the curvature information and the statistical information of the curvature; S3. Key point recognition, marking the points where the acceleration exceeds the acceleration threshold as acceleration key points, marking the points where the optimized curvature first derivative exceeds the optimized curvature first derivative threshold as optimized curvature first derivative key points, and identifying the start and end points of the hovering area in the trajectory through the top view intersection points to determine the key points in the hovering area; S4. Trajectory segmentation, generating a trajectory segmentation result by integrating the information of the acceleration key points, the optimized curvature first derivative key points, and the key points in the hovering area.

[0006] In the above solution, step S1 includes: Data cleaning, timestamp normalization, and coordinate system conversion are performed on multi-source trajectory data to generate time-series trajectory information, including the position and timestamp of trajectory points, and a set of time-series trajectory points is obtained.

[0007] In the above solution, in step S2, the acceleration corresponding to the trajectory point is calculated based on the trajectory information, and the curvature corresponding to the trajectory point is calculated based on the trajectory information and the acceleration, including: Calculate the speed corresponding to the trajectory point through the position and timestamp; Calculate the acceleration corresponding to the trajectory point through the speed and timestamp; Calculate the curvature corresponding to the trajectory point through the speed and acceleration corresponding to the trajectory point.

[0008] In the above solution, in step S2, the statistics of curvature-related information include: Statistical mean and standard deviation of curvature.

[0009] In the above solution, in step S3, the points with acceleration exceeding the acceleration threshold are marked as acceleration key points, including: Perform statistics on acceleration-related information through the acceleration of the trajectory point; Calculate the acceleration dynamic threshold based on the trajectory point according to the statistical acceleration-related information; Mark the points with acceleration exceeding the acceleration dynamic threshold as key points.

[0010] In the above solution, the statistics of acceleration-related information include: Statistical mean and standard deviation of the acceleration of the trajectory.

[0011] In the above solution, in step S3, the points with the optimized curvature first derivative exceeding the optimized curvature first derivative threshold are marked as optimized curvature first derivative key points, including: Perform statistics on optimized curvature first derivative-related information through the optimized curvature first derivative of the trajectory; Calculate the dynamic threshold of the optimized curvature first derivative based on the trajectory point according to the statistical information of the optimized curvature first derivative; Mark the points with the optimized curvature first derivative exceeding the optimized curvature first derivative dynamic threshold as key points.

[0012] In the above solution, the statistics of optimized curvature first derivative-related information include: Statistical mean and standard deviation of the optimized curvature first derivative.

[0013] In the above solution, in step S3, the start and end points of the hovering area in the trajectory are identified through the intersection points in the top view, and the key points of the hovering area are determined, including: Obtain the longitude and latitude information of the trajectory points and sort them in chronological order; Initialize the intersection set; Find all possible trajectory intersection points, and filter out the outermost intersection points through the timing relationship of the intersection points and add them to the intersection point set; Mark the outermost intersection points as key points, and mark the intermediate points as non-key points.

[0014] According to the second aspect embodiment of the technical solution of the present invention, a system for trajectory segmentation based on key point recognition is provided. The system is used to implement the method for trajectory segmentation based on key point recognition described in any one of the above solutions. The system includes: A data preprocessing module, configured to obtain multi-source trajectory data and generate timing trajectory information; A feature calculation module, configured to calculate the acceleration corresponding to the trajectory points according to the trajectory information, calculate the curvature corresponding to the trajectory points according to the trajectory information and the acceleration, perform statistics on curvature-related information through the curvature of the trajectory, and calculate the first-order derivative of the optimized curvature corresponding to the trajectory points according to the curvature information and the statistical information of the curvature; A key point recognition module, configured to mark the points with acceleration exceeding the acceleration threshold as acceleration key points, mark the points with the first-order derivative of the optimized curvature exceeding the first-order derivative threshold of the optimized curvature as key points of the first-order derivative of the optimized curvature, and identify the start and end points of the spiral area in the trajectory through the top view intersection points to determine the key points of the spiral area; A trajectory segmentation module, configured to generate a trajectory segmentation result by integrating the information of the acceleration key points, the key points of the first-order derivative of the optimized curvature, and the key points of the spiral area.

[0015] Advantages of the present invention: A method and system for trajectory segmentation based on key point recognition disclosed by the present invention, through a comprehensive analysis of trajectory data, uses statistical information and rule judgments of trajectory point feature variables such as curvature, speed, and acceleration to identify key points in the trajectory, and based on these key points, the trajectory data is segmented into multiple segments, and each segment represents a relatively independent motion trajectory. The present invention can accurately identify key points in the trajectory, especially maintaining high performance even in complex trajectories and high-noise data, effectively improving the efficiency and reliability of trajectory analysis, and thus having good application prospects in the fields of intelligent transportation management, unmanned aerial vehicle navigation, motion analysis, etc. Description of the Drawings

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on the structures shown in these drawings.

[0017] Figure 1Flowchart of the trajectory segmentation method based on key point recognition provided by the present invention; Figure 2 Flowcharts of each module of the trajectory segmentation method based on key point recognition provided by the present invention.

[0018] The realization, functional characteristics and advantages of the object of the present invention will be further described with reference to the embodiments and the accompanying drawings. Specific Embodiments

[0019] Here, the exemplary embodiments will be described in detail, and the examples are shown in the accompanying drawings. When the following description refers to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present disclosure as detailed in the appended claims.

[0020] The terms "first", "second", etc. in the specification and claims of the present disclosure are used to distinguish similar objects and do not necessarily describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments of the present disclosure described herein can be implemented in an order other than those illustrated or described herein.

[0021] In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that comprises a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0022] Multiple, including two or more.

[0023] And / or, it should be understood that for the term "and / or" used in the present disclosure, it is merely a description of the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B may represent: A exists alone, A and B exist simultaneously, and B exists alone.

[0024] As Figure 1 And Figure 2 As shown, an embodiment of the technical solution of the present invention provides a trajectory segmentation method based on key point recognition, which is applied to fields such as intelligent traffic management, unmanned aerial vehicle navigation, and motion analysis, and includes: S1. Data preprocessing, obtaining multi-source trajectory data and generating time-series trajectory information; Perform data cleaning, timestamp normalization, and coordinate system conversion on multi-source trajectory data to generate time-series trajectory information, including the position and timestamp of trajectory points, and obtain a set of time-series trajectory points as follows: Among them, represents the trajectory point, and there are a total of ; respectively represent the trajectory point in the direction component; represents the timestamp of the trajectory point.

[0025] Specifically, in this embodiment, a moving average filter or a median filter is used to remove abnormal points and remove noise to ensure the continuity of the trajectory; the timestamps of multi-source data are converted into a unified format; the geographical coordinates (such as longitude, latitude, and altitude) are converted into multiple formats (such as a three-dimensional coordinate system) to facilitate trajectory analysis.

[0026] S2. Feature calculation: Calculate the acceleration corresponding to the trajectory point according to the trajectory information, calculate the curvature corresponding to the trajectory point according to the trajectory information and the acceleration, perform curvature-related information statistics through the curvature of the trajectory, and calculate the first-order derivative of the optimized curvature corresponding to the trajectory point according to the curvature information and the statistical information of the curvature; In step S2, calculating the acceleration corresponding to the trajectory point according to the trajectory information and calculating the curvature corresponding to the trajectory point according to the trajectory information and the acceleration includes: Calculate the velocity corresponding to the trajectory point through the position and timestamp , and the formula is as follows: Among them, respectively represent the velocity components of the i-th trajectory point in the direction; respectively represent the trajectory point in the direction component; represents the timestamp of the trajectory point.

[0027] Calculate the acceleration corresponding to the trajectory point through the velocity and timestamp , and the formula is as follows: Among them, respectively represent the acceleration components of the i-th trajectory point in the direction; respectively represent the i + 1-th trajectory point in the The velocity component in the direction.

[0028] Calculate the curvature corresponding to the trajectory point through the velocity and acceleration corresponding to the trajectory point , and the formula is as follows: In step S2, the statistics of curvature-related information include: calculating the mean and standard deviation of the curvature; Calculate the mean curvature of the trajectory , and the formula is as follows: Calculate the standard deviation of the curvature of the trajectory , and the formula is as follows: Calculate the optimized curvature of the trajectory through the curvature information and the curvature smoothing factor , including: smoothing the curvature of each point by combining time series, and adapting the smoothing factor Adjust adaptively according to the noise level, and the formula is as follows: In this embodiment, the curvature formula is improved, and the curvature is smoothed by combining time series, so the noise interference is reduced. In the calculation formula of the smoothing factor , the general value is selected for the coefficient, and the coefficient can also be adjusted according to the actual situation.

[0029] Calculate the first-order derivative of the optimized curvature corresponding to the trajectory point through the optimized curvature and the timestamp , and the formula is as follows: S3. Key point identification: Mark the points where the acceleration exceeds the acceleration threshold as acceleration key points, mark the points where the first-order derivative of the optimized curvature exceeds the first-order derivative threshold of the optimized curvature as key points of the first-order derivative of the optimized curvature, identify the starting and ending points of the hovering area in the trajectory through the intersection points in the top view, and determine the key points of the hovering area; In step S3, marking the points where the acceleration exceeds the acceleration threshold as acceleration key points includes: Perform statistics on acceleration-related information through the acceleration of the trajectory points. Among them, the statistics of acceleration-related information include: calculating the mean acceleration and the standard deviation of the acceleration of the trajectory.

[0030] Calculate the mean acceleration of the trajectory , and the formula is as follows: Calculate the standard deviation of the acceleration of the trajectory , and the formula is as follows: Calculate the acceleration dynamic threshold based on trajectory points according to the statistically obtained acceleration-related information , and the formula is as follows: Mark the points where the acceleration exceeds the acceleration dynamic threshold as acceleration key points , and the formula is as follows: In step S3, mark the points where the optimized curvature first derivative exceeds the optimized curvature first derivative threshold as optimized curvature first derivative key points, including: Perform statistics on the information related to the optimized curvature first derivative through the optimized curvature first derivative of the trajectory. The statistics of the information related to the optimized curvature first derivative include: statistically calculating the mean and standard deviation of the optimized curvature first derivative; Calculate the mean of the optimized curvature first derivative of the trajectory , and the formula is as follows: Calculate the standard deviation of the optimized curvature of the trajectory , and the formula is as follows: Calculate the dynamic threshold of the optimized curvature first derivative based on trajectory points through the statistical information of the optimized curvature first derivative , and the formula is as follows: Mark the points where the optimized curvature first derivative exceeds the dynamic threshold of the optimized curvature first derivative as optimized curvature first derivative key points , and the formula is as follows: Since acceleration can directly reflect the speed change rate. When the speed in the trajectory changes significantly (such as from high speed to low speed or from low speed to high speed), the acceleration will reach a relatively high value. Therefore, by detecting acceleration, the motion change points in the trajectory can be effectively identified, which are the key points of the trajectory. In the uniform motion segment, the acceleration is close to zero, so it will not be marked as a key point. In contrast, although speed can reflect the high and low speed segments in the trajectory, it cannot distinguish the uniform motion. Therefore, acceleration can help filter out these uniform segments and only focus on the segments with significant motion changes. And an acceleration dynamic threshold is introduced to adapt to different scenarios, and the positions where the motion in the trajectory changes are screened by the acceleration exceeding the acceleration dynamic threshold.

[0031] Curvature is a physical quantity that describes the degree of curve bending. The greater the curvature, the greater the degree of curve bending. However, simply relying on curvature to identify key points may mislabel all points during the entire turning process as key points. In this embodiment, by examining the first derivative of curvature, points with significant curvature changes, that is, the true starting and ending points of turning, can be found more accurately, thereby reducing misjudgment. The first derivative of curvature can effectively detect the turning points of a trajectory. The curvature change during a large turning process may be relatively smooth, but the curvature change rate will increase significantly at the beginning and end of the turn. Therefore, the first derivative of curvature can help identify these starting and ending points of turning. And a dynamic threshold of the first derivative of curvature is introduced to adapt to different scenarios. The starting and ending positions of turning in the trajectory are determined by the first derivative of curvature exceeding the dynamic threshold of the first derivative of curvature.

[0032] In step S3, the starting and ending points of the hovering area in the trajectory are identified through the intersection points in the top view, and the key points of the hovering area are determined, including: Obtain the longitude and latitude information of the trajectory points and sort them in chronological order; Among them, represents the longitude of the th trajectory point; represents the

[0033] latitude of the .

[0034] Among them, represents the th line segment, and the endpoints are and .

[0035] Initialize the intersection point set; Among them, represents the set of all trajectory points, represents the set of latitudes of the trajectory points.

[0036] Find all possible trajectory intersection points, and filter out the outermost intersection points through the chronological relationship of the intersection points and add them to the intersection point set. Specifically as follows: (1) Traverse the line segment list , let the current th line segment be , where the endpoints of are and ; (2) Traverse the line segment list , and let the current j-th line segment be , where The endpoints of are and ; (3) Calculate the vector cross product: (4) If , then and There is an intersection point, go to (5).; Otherwise, go to (2)) (5) If is empty, then add The corresponding intersection point pair , go to (1); Otherwise, go to (6).

[0037] (6) If The corresponding timestamp is between The timestamp of the last intersection point pair in, go to (1); Otherwise, add The corresponding intersection point pair , go to (1).

[0038] Mark the outermost intersection points as key points, and mark the intermediate points as non-key points, as follows: Mark all the points between all the intersection point pairs in as non-key points, and mark all the intersection point pairs as key points in the hovering area : ; For the hovering segments in the trajectory, segment them by finding the intersection points of the trajectory on the top view. The intersection points indicate that the trajectory is at the same longitude and latitude position at two time points, which usually means that the trajectory hovers during this period. By identifying the intersection points, that is, the intersection points of the trajectory projected on the two-dimensional plane, the start and end points of the hovering area can be determined. In this implementation, for the case where there are multiple trajectory intersection points in the trajectory, retain the intersection points corresponding to the hovering trajectory segments that are adjacent and non-intersecting in the time dimension. And verify the 3D key points through the 2D projection intersection points to reduce false detections.

[0039] S4. Trajectory segmentation. Integrate the information of the acceleration key points, the optimized first-order derivative of curvature key points, and the hovering area key points to generate the trajectory segmentation result, as follows: Integrate all the identified key points , and sorted by timestamp: ; Trajectory segmentation: The trajectory data is segmented into multiple segments according to key points, and each segment represents a relatively independent and meaningful motion trajectory.

[0040] .

[0041] According to the second aspect embodiment of the technical solution of the present invention, a system for trajectory segmentation based on key point recognition is provided. The system is used to implement the method for trajectory segmentation based on key point recognition described in any one of the above solutions. The system includes: A data preprocessing module, configured to obtain multi-source trajectory data and generate time-series trajectory information; A feature calculation module, configured to calculate the acceleration corresponding to the trajectory points according to the trajectory information, calculate the curvature corresponding to the trajectory points according to the trajectory information and the acceleration, perform statistics on curvature-related information through the curvature of the trajectory, and calculate the first-order derivative of the optimized curvature corresponding to the trajectory points through the curvature information and the statistical information of the curvature; A key point recognition module, configured to mark the points where the acceleration exceeds the acceleration threshold as acceleration key points, mark the points where the first-order derivative of the optimized curvature exceeds the first-order derivative threshold of the optimized curvature as key points of the first-order derivative of the optimized curvature, and identify the start and end points of the hovering area in the trajectory through the top view intersection points to determine the key points of the hovering area; A trajectory segmentation module, configured to generate a trajectory segmentation result by integrating the information of the acceleration key points, the key points of the first-order derivative of the optimized curvature, and the key points of the hovering area.

[0042] It should be noted that in this article, the terms "include", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device. Without further limitation, the element defined by the statement "including a..." does not exclude the existence of additional identical elements in the process, method, article or device including the element.

[0043] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages and disadvantages of the embodiments.

[0044] Through the description of the above embodiments, those skilled in the art can clearly understand that the above implementation methods can be realized by means of software plus a necessary general hardware platform. Of course, it can also be realized by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes several instructions for causing a terminal (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in various embodiments of the present invention.

[0045] The embodiments of the present invention have been described above in conjunction with the accompanying drawings. However, the present invention is not limited to the above specific implementation manners. The above specific implementation manners are merely illustrative rather than restrictive. Under the inspiration of the present invention, those of ordinary skill in the art can also make many forms without departing from the purpose of the present invention and the scope protected by the claims. All of these are within the protection scope of the present invention.

Claims

1. A trajectory segmentation method based on key point recognition, characterized in that Including: S1. Data preprocessing, obtaining multi-source trajectory data and generating time-series trajectory information; S2. Feature calculation, calculating the acceleration corresponding to the trajectory points according to the trajectory information, calculating the curvature corresponding to the trajectory points according to the trajectory information and the acceleration, performing statistics on curvature-related information through the curvature of the trajectory, and calculating the first-order derivative of the optimized curvature corresponding to the trajectory points according to the curvature information and the statistical information of the curvature; S3. Key point identification, marking the points where the acceleration exceeds the acceleration threshold as acceleration key points, marking the points where the first-order derivative of the optimized curvature exceeds the first-order derivative threshold of the optimized curvature as key points of the first-order derivative of the optimized curvature, and identifying the start and end points of the spiral area in the trajectory through the top-view intersection points to determine the key points of the spiral area; S4. Trajectory segmentation, generating the trajectory segmentation result by integrating the information of the acceleration key points, the key points of the first-order derivative of the optimized curvature, and the key points of the spiral area.

2. The trajectory segmentation method based on key point recognition according to claim 1, wherein Step S1 includes: Performing data cleaning, timestamp normalization, and coordinate system conversion on the multi-source trajectory data to generate time-series trajectory information. The trajectory information includes the position and timestamp of the trajectory points, and a set of time-series trajectory points is obtained.

3. The trajectory segmentation method based on key point recognition according to claim 2, wherein, In step S2, calculating the acceleration corresponding to the trajectory points according to the trajectory information, and calculating the curvature corresponding to the trajectory points according to the trajectory information and the acceleration, including: Calculating the speed corresponding to the trajectory points through the position and timestamp; Calculating the acceleration corresponding to the trajectory points through the speed and timestamp; Calculating the curvature corresponding to the trajectory points through the speed and acceleration corresponding to the trajectory points.

4. The method for segmenting a trajectory based on key point recognition according to claim 1, wherein In step S2, the statistics of the curvature-related information include: Statistical mean and standard deviation of the curvature.

5. The trajectory segmentation method based on key point recognition according to claim 1, characterized in that In step S3, marking the points where the acceleration exceeds the acceleration threshold as acceleration key points, including: Performing statistics on acceleration-related information through the acceleration of the trajectory points; Calculating the dynamic threshold of the acceleration based on the trajectory points according to the statistical acceleration-related information; Marking the points where the acceleration exceeds the dynamic threshold of the acceleration as key points.

6. The trajectory segmentation method based on key point recognition according to claim 5, wherein, The statistics of the acceleration-related information include: Statistical mean and standard deviation of the acceleration of the trajectory.

7. The trajectory segmentation method based on key point recognition according to claim 1, wherein In step S3, marking the points where the first-order derivative of the optimized curvature exceeds the first-order derivative threshold of the optimized curvature as key points of the first-order derivative of the optimized curvature, including: Performing statistics on the first-order derivative of the optimized curvature-related information through the first-order derivative of the optimized curvature of the trajectory; Calculating the dynamic threshold of the first-order derivative of the optimized curvature based on the trajectory points through the statistical information of the first-order derivative of the optimized curvature; Marking the points where the first-order derivative of the optimized curvature exceeds the dynamic threshold of the first-order derivative of the optimized curvature as key points.

8. The method for segmenting a trajectory based on key point recognition according to claim 7, wherein, The statistics of the first-order derivative of the optimized curvature-related information include: Statistical mean and standard deviation of the first-order derivative of the optimized curvature.

9. The method for segmenting a trajectory based on key point recognition according to claim 1, wherein, In step S3, identifying the start and end points of the spiral area in the trajectory through the top-view intersection points to determine the key points of the spiral area, including: Obtaining the longitude and latitude information of the trajectory points and sorting them in chronological order; Initializing the intersection set; Finding all possible trajectory intersections, and screening out the outermost intersections through the chronological relationship of the intersections and adding them to the intersection set; Marking the outermost intersections as key points and marking the intermediate points as non-key points.

10. A system for trajectory segmentation based on key point recognition, characterized in that, The system is used to implement the trajectory segmentation method based on key point identification described in any one of claims 1-9. The system includes: A data preprocessing module, which is used to obtain multi-source trajectory data and generate temporal trajectory information; A feature calculation module, which is used to calculate the acceleration corresponding to a trajectory point according to the trajectory information, calculate the curvature corresponding to the trajectory point according to the trajectory information and the acceleration, perform curvature-related information statistics through the curvature of the trajectory, and calculate the optimized curvature first derivative corresponding to the trajectory point through the curvature information and the statistical information of the curvature; A key point identification module, which is used to mark the points where the acceleration exceeds the acceleration threshold as acceleration key points, mark the points where the optimized curvature first derivative exceeds the optimized curvature first derivative threshold as optimized curvature first derivative key points, identify the start and end points of the hovering area in the trajectory through the top view intersection points, and determine the key points of the hovering area; A trajectory segmentation module, which is used to generate a trajectory segmentation result by integrating the information of acceleration key points, optimized curvature first derivative key points and key points of the hovering area.

Images