A trajectory spatio-temporal co-occurrence pattern visual analysis method

Abstract

The application provides a trajectory spatiotemporal co-occurrence pattern visual analysis method, comprising segmenting trajectory data to obtain trajectory routes, calculating correlation intensity between the trajectory routes according to co-occurrence matching values, and dividing the trajectory routes into path clusters according to the correlation intensity; obtaining N-level combinations in which spatiotemporal co-occurrence intensity values in the path clusters are higher than a threshold value through pruning optimization; and obtaining related data of the N-level combinations to generate several visual projections which can be information-interacted. The application can quickly and efficiently reduce the large data scale of spatiotemporal trajectory data, extract important spatiotemporal information in the trajectory, and use visual views and interactive modes to facilitate intuitive and visual understanding of the data.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and more specifically to a visual analysis method for trajectory spatiotemporal co-occurrence patterns. Background Technology

[0002] The position of an object relative to the Earth is generally measured using geospatial data, namely the three dimensions of longitude, latitude, and altitude. In daily life, the development of things is often accompanied by the passage of time, in addition to changes in geographical location. Therefore, spatiotemporal data, which combines time and spatial data, has gradually become one of the main representatives of geospatial data.

[0003] With the development of artificial satellite technology and the widespread use of mobile and positioning devices, the means of collecting spatiotemporal information are constantly being enriched and developed, and massive amounts of spatiotemporal data are being generated all the time. Spatiotemporal data is gradually showing characteristics of being multi-source, heterogeneous, large-scale, multi-semantic, and multimodal. However, the results of time and space data analysis are not easy to interpret and understand. After acquiring massive amounts of spatiotemporal data, it is not possible to quickly understand the relationship between different spatiotemporal data.

[0004] Spatiotemporal co-occurrence patterns, as the name suggests, refer to objects being close to each other in both time and space. Specifically, given a dataset containing multiple moving objects, a spatiotemporal co-occurrence pattern refers to a subset of object types whose instances are spatially close over a long time span.

[0005] Spatiotemporal trajectory data includes personal trajectories, which are relatively personalized and continuous, accurately reflecting an individual's lifestyle and behavioral characteristics. The spatiotemporal co-occurrence patterns of personal trajectories also have significant research value: they can help identify anomalous events and encounters between suspicious individuals and others; for predicting the spread and infection of infectious diseases, spatiotemporal co-occurrence patterns can help analyze the spread and development of infectious diseases.

[0006] However, processing the spatiotemporal co-occurrence of personal trajectories presents significant challenges due to the large volume of data required, which severely impacts processing efficiency. Personal GPS trajectory data itself is characterized by densely distributed sampling points and a massive scale, making data processing and mining analysis extremely complex, with only a small portion potentially being useful for trajectory analysis. Therefore, reducing the complexity of mining methods, effectively and quickly extracting and analyzing relevant information from personal trajectory data, and visually representing spatiotemporal co-occurrence patterns have become major challenges. Summary of the Invention

[0007] In view of this, the problem to be solved by the present invention is to provide a visual analysis method for the spatiotemporal co-occurrence pattern of trajectory, which can quickly and efficiently reduce the massive data scale of spatiotemporal trajectory data, extract important spatiotemporal information in the trajectory, and use visualization view and interactive method to facilitate people's intuitive and vivid understanding of the data.

[0008] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: A visual analysis method for trajectory spatiotemporal co-occurrence patterns includes: Step 1: Segment the trajectory data based on the dwell segments to obtain the trajectory routes. Calculate the correlation strength between trajectory routes based on the co-occurrence matching values ​​among all dwell segments of different trajectory routes. Then, cluster the trajectory data based on the correlation strength to generate several path clusters. Step 2: Obtain several trajectory routes within the path cluster and combine them in pairs to generate a first-level combination. Calculate the spatiotemporal co-occurrence intensity value of the first-level combination and remove the first-level combination whose spatiotemporal co-occurrence intensity value is lower than a preset threshold. Combine the trajectory routes in groups of three to generate several second-level combinations. After removing the second-level combinations that contain the removed first-level combinations, calculate the spatiotemporal co-occurrence intensity value of the second-level combinations and remove the second-level combinations whose spatiotemporal co-occurrence intensity value is lower than a preset threshold. Repeat the above steps to generate several N-level combinations in which the spatiotemporal co-occurrence intensity values ​​are all higher than the threshold. Step 3: Obtain the trajectory routes, trajectory data, and correlation strength of the N-level combinations to generate several interactive visualization projections.

[0009] Furthermore, the method for obtaining personal trajectory data involves: acquiring trajectory data and extracting dwell segments whose dwell time exceeds a preset time threshold along the time sequence; connecting the dwell segments sequentially in the time sequence to generate the person's trajectory route.

[0010] Furthermore, the co-occurrence matching value includes determining whether two dwelling segments co-occur in time and space. The determination method is as follows: the dwelling segments are respectively represented as period1 and period12, the time threshold is θt, and the distance threshold is θd.

[0011]

[0012] overlap(period1, period2) is the duration of overlap between two trajectory segments in the time domain, and dis(period1, period2) is the spatial distance between the two trajectory segments. When both the spatial distance and the duration of time meet the threshold, the two dwell segments co-occur in time and space, and the two dwell segments are defined as a dwell segment group.

[0013] Furthermore, the method for extracting all the stop segment groups between the trajectory routes is as follows: scan along the time axis of the trajectory route, and each time move the index of the stop segment with the earlier end time to the back one stop segment, until one of the trajectories is finally traversed, so as to quickly complete the indexing of all stop segments on two trajectories and filter out all the stop segment groups between the trajectory routes.

[0014] Furthermore, the method for obtaining the correlation between the two trajectory routes is as follows: All stop segment groups between the trajectory routes are obtained, and the temporal correlation, spatial correlation, and spatiotemporal correlation of each stop segment group are calculated. The calculation formulas are as follows:

[0015]

[0016]

[0017] The temporal correlation value includes: obtaining the ratio of the intersection to the union of the two dwelling periods in the time domain to generate the Jaccard correlation coefficient, and exponentializing the Jaccard correlation coefficient; the spatial correlation value includes: the distance between the two dwelling periods divided by the spatial distance threshold θd, and exponentializing this ratio; the spatiotemporal correlation value is obtained by multiplying the temporal correlation measure and the spatial correlation measure. The average of all spatiotemporal correlation values ​​is defined as the spatiotemporal co-occurrence intensity value between two trajectory routes.

[0018] Furthermore, the clustering process can use K-Means clustering algorithm, DBSCAN clustering algorithm or spectral clustering algorithm to aggregate trajectory data with similar movement trajectories into a path group.

[0019] Furthermore, the visualization projection includes a map view that can display trajectory paths, partition layers, and heat map layers; The partition layer divides the map view into regions based on the dwell time segment and marks them with different colors; the heat map layer is used to display high-density areas where spatiotemporal co-occurrence occurs.

[0020] Furthermore, the visualization projection includes a co-occurrence pattern diagram with the horizontal axis representing time, where the horizontal axis corresponds to a vertical box representing a stop segment, and different trajectory routes are sequentially connected to the corresponding vertical boxes according to the time corresponding to the stop segment; The color of the vertical frame corresponds to the color of the pause segment indicated by the partition layer, the height of the region frame corresponds to the number of co-occurring trajectories, and the transparency of the region frame corresponds to the spatiotemporal co-occurrence intensity.

[0021] Furthermore, the visualization projection includes a trajectory view, which includes several vertically arranged horizontal frames. Color blocks are arranged and filled within the horizontal frames. The color of the color block corresponds to the color of the pause segment in the partition layer, and the length of the color block corresponds to the time period.

[0022] Furthermore, the visualization projection includes a correlation matrix view with a square frame structure, where both the horizontal and vertical axes are used to represent the trajectory route, and the squares at the intersection of the horizontal and vertical axes correspond to the correlation between the horizontal and vertical axes. The higher the correlation between the trajectory routes, the darker the color of the box at the corresponding intersection point.

[0023] The advantages and positive effects of this invention are: 1. By segmenting the trajectory based on stop points, and then performing correlation analysis and clustering based on trajectory segment matching to classify and filter the trajectories, the spatiotemporal co-occurrence patterns among a certain number of specific trajectories are finally extracted. This reduces the massive data size of spatiotemporal trajectory data and quickly extracts important spatiotemporal information from the trajectories.

[0024] 2. By using storyline visualization to represent spatiotemporal co-occurrence patterns and providing various interactive methods between visualization projections, a visual exploration is facilitated. Based on this, various visualization views and interactive methods are developed around the geospatial information, correlation analysis, and spatiotemporal co-occurrence patterns of personal trajectory data, forming a visual analysis system to help users explore and analyze. This allows people to intuitively and visually understand the data, enabling the rapid identification of relevant individuals and addresses in trajectory analysis, prediction and control of infectious disease transmission. Attached Figure Description

[0025] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is an overall flowchart of a visual analysis method for trajectory spatiotemporal co-occurrence patterns according to the present invention; Figure 2 This is a conceptual diagram of trajectory segment matching in a visual analysis method for trajectory spatiotemporal co-occurrence patterns according to the present invention; Figure 3 This is a schematic diagram of the time-axis scanning optimization of a visual analysis method for trajectory spatiotemporal co-occurrence patterns according to the present invention; Figure 4 This is a correlation matrix view of a visual analysis method for trajectory spatiotemporal co-occurrence patterns according to the present invention; Figure 5 This is a pruning optimization diagram of a visual analysis method for trajectory spatiotemporal co-occurrence patterns according to the present invention; Figure 6 This is a map view of a visual analysis method for trajectory spatiotemporal co-occurrence patterns according to the present invention; Figure 7 This is a co-occurrence pattern diagram of a visual analysis method for trajectory spatiotemporal co-occurrence patterns according to the present invention; Figure 8 This is a trajectory view of a visual analysis method for trajectory spatiotemporal co-occurrence patterns according to the present invention; Figure 9 This is the overall display boundary map of the visualization projection of a visual analysis method for a trajectory spatiotemporal co-occurrence mode according to the present invention. Detailed Implementation

[0026] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0027] It should be noted that when a component is described as "fixed to" another component, it can be directly on the other component or may have a component in between. When a component is described as "connected to" another component, it can be directly connected to the other component or may have a component in between. When a component is described as "set on" another component, it can be directly set on the other component or may have a component in between. The terms "vertical," "horizontal," "left," "right," and similar expressions used in this document are for illustrative purposes only.

[0028] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0029] This invention provides a visual analysis method for trajectory spatiotemporal co-occurrence patterns, such as... Figure 1 As shown, taking the processing of data on the activity trajectory of a suspect as an example, the trajectory data of several suspects are obtained, as well as the trajectory data of several individuals within the area where the suspect stayed during the same period of the suspect's movement.

[0030] Step 1: Segment the trajectory data based on the stop segments to obtain the trajectory routes. Calculate the correlation strength between trajectory routes based on the co-occurrence matching values ​​among all stop segments of different trajectory routes. Then, cluster the trajectory data based on the correlation strength to generate several path clusters.

[0031] Human trajectory data is usually located by GPS. Because GPS trajectory data is generally sampled at dense points, with sampling and recording almost every second, the amount of data is extremely large. In order to save storage and data processing time, trajectory data is preprocessed to generate trajectory routes.

[0032] The process of generating a trajectory route is as follows: Extract a trajectory data, read the information of the trajectory data in chronological order, and when it is detected that a person stays at a certain address for a period of time exceeding a preset time threshold, extract the trajectory data segment (from entering the address to leaving the address) and generate a stop segment. Continue to read the trajectory data after the trajectory data segment until all stop segments are extracted. All stop segments are connected in chronological order to generate the trajectory route of this person.

[0033] After preprocessing all the trajectory data and generating trajectory data, the trajectory correlation between the trajectories is calculated, and the trajectories are classified and aggregated into several path clusters based on the correlation between the trajectory routes.

[0034] The calculation process for trajectory correlation includes: (1) Determine whether there is spatiotemporal co-occurrence between the stopping segments on different trajectory routes.

[0035] like Figure 2 As shown, the calculation process is as follows: Arbitrarily combine two trajectory routes, and set two stop segments on each of the two trajectory routes, denoted as period1 and period12 respectively (corresponding to...). Figure 2 The trajectory stay segment e A and trajectory stay segment e B The time threshold is θt, and the distance threshold is θd;

[0036]

[0037] overlap(period1, period2) is the duration of overlap between two trajectory segments in the time domain, and dis(period1, period2) is the spatial distance between the two trajectory segments. When both of the above values ​​meet the requirements, the two trajectory points are considered to be matched in time and space.

[0038] like Figure 3As shown, since the trajectory route has a certain time orientation, by scanning along the time axis, the index of the stop segment corresponding to the earlier end time is moved forward by one stop segment each time, until one of the trajectories is finally traversed, thus completing the indexing of all stop segments on both trajectories in the fastest way. This quickly and efficiently finds groups of stop segments that coexist in time and space.

[0039] There's no need to arbitrarily match every pair of stop segments on two trajectory routes and then determine the spatiotemporal co-occurrence between them, greatly reducing the number of matching operations and the algorithm's complexity. For example, if one trajectory route has M stop segments and another has N stop segments, the combinations of stop segments between the two trajectory routes are M×N. When scanning along the time axis, the number of matching combinations does not exceed M+N.

[0040] Using the same method, all trajectory routes are combined in pairs to calculate the dwell time segments where the two trajectory routes coexist in time.

[0041] (2) Determine the time co-occurrence of the two trajectory routes and calculate the time correlation, spatial correlation and spatiotemporal correlation between the time co-occurrence groups.

[0042] The method for calculating time correlation is as follows: The Jaccard correlation coefficient is generated by the ratio of the intersection to the union of the two time domains, and then the Jaccard correlation coefficient is indexed. The specific formula is as follows:

[0043] Spatial correlation is calculated by dividing the distance between two stops by a spatial distance threshold θd, and then exponentializing this ratio. The specific formula is as follows:

[0044] The spatiotemporal correlation is calculated by multiplying the temporal correlation measure and the spatial correlation measure of two dwell periods. The specific formula is as follows:

[0045] (3) After calculating the temporal correlation, spatial correlation and spatiotemporal correlation of all the stop segments between the two trajectory routes, calculate the average value of the temporal-spatial correlation to define the correlation strength between different trajectory routes.

[0046] A correlation matrix view is constructed based on the correlation strength. This view is used for visualization projection to intuitively understand the correlation strength between different trajectory routes.

[0047] The correlation matrix view is a square box diagram. Both the horizontal and vertical axes represent the trajectory routes. The boxes at the intersection of the horizontal and vertical axes correspond to the correlation between the trajectories on the corresponding axes. For example... Figure 4 As shown, the darker the color inside the box, the higher the correlation between the trajectories. A white box indicates that they are the same trajectory (to avoid affecting the judgment). Preferably, to facilitate the separate judgment of the correlation between multiple trajectories, the boxes with high correlation are placed close to the diagonal of the square diagram.

[0048] Based on the correlation strength between different trajectory routes, clustering algorithms are more practical to group trajectory routes into several path clusters, making the trajectory routes within each path cluster similar. Clustering algorithms include K-Means clustering, DBSCAN clustering, and spectral clustering.

[0049] Step 2: Combine several trajectory routes within the path cluster in pairs to generate a first-level combination and calculate the spatiotemporal co-occurrence intensity value of the combination. Remove the first-level combinations whose spatiotemporal co-occurrence intensity values ​​are lower than a preset threshold. Combine personal trajectory data in groups of three to generate a second-level combination. After removing the second-level combinations that contain the removed first-level combinations, calculate the spatiotemporal co-occurrence intensity value of each second-level combination. Remove the second-level combinations whose spatiotemporal co-occurrence intensity values ​​are lower than a preset threshold. Repeat the above steps to generate several N-level combinations in which all spatiotemporal co-occurrence intensity values ​​are higher than the threshold.

[0050] Based on the combination of path clusters, several trajectory routes within each path cluster are extracted. The example extracts several trajectory routes of suspects and several trajectory routes with a correlation higher than a set value. The number of trajectory routes and the correlation strength can be set manually. Suspect trajectory routes and several trajectory routes with high correlation to suspects can be selected. The correlation level can be viewed in the correlation matrix view.

[0051] like Figure 5 As shown, the trajectory routes are combined in pairs to generate first-level combinations, and the spatiotemporal co-occurrence intensity value of each combination is calculated. After removing first-level combinations whose spatiotemporal co-occurrence intensity values ​​are below a preset threshold, the individual trajectory data are combined in groups of three to generate second-level combinations. After removing second-level combinations that contain the removed first-level combinations, the spatiotemporal co-occurrence intensity value of each second-level combination is calculated again, and second-level combinations whose spatiotemporal co-occurrence intensity values ​​are below a preset threshold are removed. This process is repeated to generate several N-level combinations where all spatiotemporal co-occurrence intensity values ​​are above the threshold. If X trajectory routes are extracted, N <X-1。

[0052] Utilizing the prior property concept in the Apriori algorithm—that all non-empty subsets of a frequent itemset are frequent, and all non-empty parent sets of an infrequent itemset are infrequent—combinations with low relevance can be removed promptly, improving data processing efficiency. The calculation method for the spatiotemporal co-occurrence intensity of N-level combinations is the same as that for two trajectory routes. Specifically, the process involves identifying all groups of stop segments (each group containing N stop segments) that co-occur in the N trajectory routes, calculating the spatiotemporal co-occurrence intensity of these groups, and then averaging the results.

[0053] Step 3: Obtain the trajectory routes of N-level combinations, the co-occurrence dwell segments between different trajectory routes, the dwell time range of the dwell segments, and the strength of co-occurrence between trajectories, and generate a visualization projection.

[0054] The visualization projection includes a map view, which contains all the stops in the N-level combination of the trajectory route, and is used to display the trajectory route.

[0055] like Figure 6 As shown, the map view has a partition layer, which can divide the map view into regions based on the dwell time segment to generate dwell time segment areas, and mark different dwell time segment areas with different colors (polygonal areas in the figure). The map view also has a heat map layer, which can be used to visualize high-density areas of spatiotemporal co-occurrence in geographic data (dotted areas in the figure).

[0056] like Figure 7 As shown, the visualization projection includes a co-occurrence pattern diagram. The horizontal axis of the co-occurrence pattern diagram corresponds to time. Different time periods correspond to the dwelling segments that produce spatiotemporal co-occurrence. The dwelling segments are vertical rectangular box structures, and the color of the rectangular box corresponds to the color of the dwelling segment area on the map view.

[0057] Different trajectory routes are connected to corresponding rectangular boxes in sequence according to the dwell period and the corresponding time. The height of the rectangular box corresponds to the number of spatiotemporal co-occurrences of the co-occurring trajectory routes, and the transparency of the rectangular box corresponds to the correlation strength.

[0058] It is mainly used to reflect the spatiotemporal co-occurrence and intensity of different trajectory routes. The number of trajectory routes can also be specified manually.

[0059] like Figure 8 As shown, the visualization projection includes a trajectory view, which consists of horizontal rectangular frames. Each rectangle corresponds to a trajectory route, and the frames are filled with color blocks. The color of the color block corresponds to the color of the pause area, and the length of the color block corresponds to the time period. This clearly and simply demonstrates the length of the overlap time of the pauses for different trajectory routes.

[0060] When displaying, several fixed trajectory routes can be specified and displayed as a trajectory view, with the relevance decreasing from top to bottom. This is used to display the overlap time length of dwell segments in different trajectory routes.

[0061] The visualization projection also includes a correlation matrix view, which is used to display the correlation between different trajectory routes.

[0062] like Figure 9 As shown, the overall display interface of the visualization projection includes a related matrix view, a trajectory view, a co-occurrence pattern diagram, and a map view. When the number and types of trajectory routes in one view are adjusted, the number and types of trajectory routes in other views change synchronously, enabling interaction between multiple visualization views for multi-angle data filtering.

[0063] The embodiments of the present invention have been described in detail above, but the content described is only a preferred embodiment of the present invention and should not be considered as limiting the scope of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of this patent.

Claims

1. A visual analysis method for trajectory spatiotemporal co-occurrence patterns, characterized in that, Including: Step 1: Segment the trajectory data based on the dwell time segments to obtain the trajectory route. The method for obtaining personal trajectory data is as follows: acquire the trajectory data and extract the dwell time segments that exceed the preset time threshold in chronological order. Connect the dwell time segments in chronological order to generate the person's trajectory route. The correlation strength between different trajectory routes is calculated based on the co-occurrence matching value among all the stops of different trajectory routes. Based on the correlation strength, the trajectory data is clustered to generate several path clusters. The co-occurrence matching value includes a method for determining whether two dwelling segments co-occur in time and space. The method for determining this is as follows: the dwelling segments are respectively represented as period1 and period12, the time threshold is θt, and the distance threshold is θd. overlap(period1, period2) is the duration of overlap between two trajectory segments in the time domain, and dis(period1, period2) is the spatial distance between the two trajectory segments. When both the spatial distance and the duration of time meet the threshold, the two dwelling segments co-occur in time and space, and the two dwelling segments are defined as a dwelling segment group. The method to extract all the stop segment groups between the trajectory routes is as follows: scan along the time axis of the trajectory route, and move the index of the stop segment with the earlier end time to the back one stop segment each time, until one of the trajectories is finally traversed, so as to quickly complete the indexing of all stop segments on two trajectories and filter out all the stop segment groups between the trajectory routes. The method for obtaining the correlation between the two trajectory routes is as follows: All stop segment groups between the trajectory routes are obtained, and the temporal correlation, spatial correlation, and spatiotemporal correlation of each stop segment group are calculated. The calculation formulas are as follows: The temporal correlation value includes: obtaining the ratio of the intersection to the union of the two stay periods in the time domain to generate the Jaccard correlation coefficient, and then exponentializing the Jaccard correlation coefficient; the spatial correlation value includes: the distance between the two stay periods divided by a spatial distance threshold θ. d This ratio is then indexed; the spatiotemporal correlation value is obtained by multiplying the temporal correlation measure and the spatial correlation measure. The average of all spatiotemporal correlation values ​​is defined as the spatiotemporal co-occurrence intensity value between two trajectory routes; Step 2: Obtain several trajectory routes within the path cluster and combine them in pairs to generate a first-level combination. Calculate the spatiotemporal co-occurrence intensity value of the first-level combination and remove the first-level combination whose spatiotemporal co-occurrence intensity value is lower than a preset threshold. Combine the trajectory routes in groups of three to generate several second-level combinations. After removing the second-level combinations that contain the removed first-level combinations, calculate the spatiotemporal co-occurrence intensity value of the second-level combinations and remove the second-level combinations whose spatiotemporal co-occurrence intensity value is lower than a preset threshold. Repeat the above steps to generate several N-level combinations in which the spatiotemporal co-occurrence intensity values ​​are all higher than the threshold. Step 3: Obtain the trajectory routes, trajectory data, and correlation strength of the N-level combinations to generate several interactive visualization projections.

2. The visual analysis method for trajectory spatiotemporal co-occurrence patterns according to claim 1, characterized in that, The clustering process can use K-Means clustering algorithm, DBSCAN clustering algorithm or spectral clustering algorithm to aggregate trajectory data with similar movement trajectories into a path group.

3. The visual analysis method for trajectory spatiotemporal co-occurrence patterns according to claim 1, characterized in that, The visualization projection includes a map view that can display trajectory paths, partition layers, and heat map layers; The partition layer divides the map view into regions based on the dwell time segment and marks them with different colors; the heat map layer is used to display high-density areas where spatiotemporal co-occurrence occurs.

4. The visual analysis method for trajectory spatiotemporal co-occurrence patterns according to claim 1, characterized in that, The visualization projection includes a co-occurrence pattern diagram with time as the horizontal axis, and the horizontal axis corresponds to the vertical box representing the dwell segment. Different trajectory routes are connected to the corresponding vertical boxes in sequence according to the time corresponding to the dwell segment. The color of the vertical frame corresponds to the color of the segment indicated by the partition layer, the height of the vertical frame corresponds to the number of co-occurring trajectories, and the transparency of the vertical frame corresponds to the spatiotemporal co-occurrence intensity.

5. A visual analysis method for trajectory spatiotemporal co-occurrence patterns according to claim 1, characterized in that, The visualization projection includes a trajectory view, which has several vertically arranged horizontal frames. Color blocks are arranged and filled in the horizontal frames. The color of the color block corresponds to the color of the pause segment in the partition layer, and the length of the color block corresponds to the time period.

6. The visual analysis method for trajectory spatiotemporal co-occurrence patterns according to claim 1, characterized in that, The visualization projection includes a related matrix view with a square box diagram structure. Both the horizontal and vertical axes are used to represent the trajectory route. The squares at the intersection of the horizontal and vertical axes correspond to the correlation between the horizontal and vertical axes. The higher the correlation between the trajectory routes, the darker the color of the box at the corresponding intersection point.

Images