A spatiotemporal trajectory-based POI visit probability measurement method and system

Abstract

The application discloses a POI visiting probability measurement method and system based on space-time trajectory, and the method comprises the following steps: acquiring the space-time trajectory of pedestrians in a target area and generating a space-time trajectory point sequence; constructing the potential visiting space-time range of the pedestrians by using the principle of time geography, and screening POI points in the potential visiting space-time range; calculating the position probability of the POI points in the potential visiting space-time range being visited; calculating the time probability of the POI points being visited; and fusing the position probability and the time probability of the POI points being visited, so as to obtain the probability of the POI points in the potential visiting space-time range being visited. According to the principle of time geography, the uncertainty space-time position between the known space-time trajectory points of the pedestrians is calculated, the POI points in the potential visiting space-time range of the pedestrians are screened, the time preference of different types of POI being visited in different time periods is weighted, the space information and the time information of the POI are fused, the space-time information of the POI being visited is expressed, and the expression of the probability measurement is more accurate.

Description

Technical Field

[0001] The present invention belongs to the technical field of spatiotemporal data mining, and in particular relates to a method and system for measuring POI visit probability based on spatiotemporal trajectory. Background Art

[0002] With the development of modern positioning technology and the widespread adoption of mobile devices, pedestrian and vehicle trajectory data has become increasingly accessible. Spatiotemporal trajectory data is widely used in fields such as tourism, transportation, and GIS. Time geography focuses on the spatial and temporal attributes of trajectories. A spatiotemporal path connects two temporally adjacent known trajectory points using a straight line, expressing the movement trend of an individual and containing both spatial location information and temporal semantic information.

[0003] When people choose activities, they tend to prefer different types of POIs based on their planned activity times. In urban spaces, spatiotemporal trajectory data is combined with POI data. By analyzing large amounts of spatiotemporal trajectory data, we can determine the frequency of individuals visiting various types of POIs during specific time periods. Counting the frequency of POI visits is one of the methods for analyzing urban hotspots.

[0004] Temporal geography has been used to measure the uncertainty of spatiotemporal trajectories. The primary tool is the spacetime prism, which represents the area a mobile object can reach, subject to the constraints of its starting point, destination, time, maximum speed, and geographic environment. This is the set of all possible trajectories. Given the readily available trajectory data, how to use spatiotemporal trajectories to measure the probability of a point of interest (POI) being visited is a key focus of temporal geography.

[0005] Currently, the calculation of a location's visit probability is largely based on spatial factors, such as distance decay. CN107704517A discloses a method for recommending points of interest based on local trajectories. This method counts the user's movement trajectories within the local activity area, calculates the user's movement probability at different distances, sorts candidate locations based on their check-in probability, and returns the candidate with the highest probability. However, the actual visit probability depends not only on spatial factors but also on non-spatial factors, such as time-related preferences.

[0006] Furthermore, spatiotemporal trajectory data often only contains information about the known anchor points of an individual moving object's trajectory, while ignoring the large amount of uncertain location information between these anchor points. For example, when using mobile devices to locate location points, there is typically a certain sampling interval, which can lead to omissions of some location information. Furthermore, when obtaining individual spatiotemporal trajectory data through interviews or questionnaires, the data can also be incomplete due to the fading of human memory. Summary of the Invention

[0007] In view of this, the present invention proposes a POI visit probability measurement method and system based on spatiotemporal trajectory, which is used to solve the problem that the existing calculation of the probability of a location being visited ignores a large amount of uncertain location point information between anchor points.

[0008] In a first aspect, the present invention discloses a method for measuring POI visit probability based on spatiotemporal trajectories, the method comprising:

[0009] Obtain the spatiotemporal trajectory of pedestrians in the target area and generate a spatiotemporal trajectory point sequence;

[0010] Using the principles of time geography, we construct the potential space-time range of pedestrians, generate the pedestrian's space-time prism and potential path area, and use the potential path area to filter out POI points within the potential space-time range.

[0011] The inverse distance weighted method is used to calculate the probability of POI points being visited within the potential space-time range.

[0012] Assign a time weight to each POI type, and obtain the time weight corresponding to the POI points within the potential access time and space range as the time probability of the POI point being visited;

[0013] The location probability and time probability of POI points being visited are integrated to obtain the probability of POI points being visited within the potential visiting time and space range.

[0014] On the basis of the above technical solution, preferably, the space-time trajectory is a time series of spatial positions, and in the space-time trajectory point sequence, each space-time trajectory point includes the spatial position and corresponding time of a trajectory point passed by the pedestrian.

[0015] Based on the above technical solutions, preferably, the method of using time geography principles to construct the potential space-time range of pedestrian visits and generate pedestrian space-time prisms and potential path areas specifically includes:

[0016] The i-th trajectory point A i The subsequent trajectory point in the tense is recorded as A j , that is, A i 、A j are two temporal neighbor space-time trajectory point pairs, A i 、A j The corresponding times are t i , t j ;

[0017] Based on the pedestrian's maximum movement speed, the principle of time geography is used to construct the potential space-time range of pedestrians' visits within two space-time trajectory points of temporal neighbors:

[0018]

[0019] Where z ijt Indicates A i 、A j The space-time prism with anchor points as vertices represents the potential space-time range of pedestrians; p represents a three-dimensional space-time point, i is the number of the space-time trajectory point, i = 0, 1, 2, ..., n, j = i + 1, j ≤ n, the total number of space-time trajectory points is n + 1, |||·||| represents the Euclidean distance in three-dimensional space, and v max is the maximum moving speed of pedestrians;

[0020] The footprint or projection of the space-time prism in the plane space serves as the potential path area for pedestrians:

[0021] g ij ={p′||p′-A′ i ||+||A′ j -p′||≤(t i -t j )·v max}

[0022] Where g ij It is prism z ijt The projection in the plane space represents the potential path area of ​​the pedestrian, ||·|| represents the Euclidean distance in the two-dimensional space, p′ represents the position point in the plane space, A′ i , A′ j A i 、A j Projection in plane space.

[0023] Based on the above technical solution, preferably, the method of screening out POI points within the potential space-time range of pedestrians by using the potential path area specifically includes:

[0024] The set of all POI points in the target area and the potential path area g ij The intersection of , obtains the POI points within the potential space-time range of pedestrians:

[0025] S ij =S∩g ij

[0026] Where S ij is the set of POI points within the potential space-time range visited by pedestrians, and S is the set of all POI points in the target area.

[0027] On the basis of the above technical solution, preferably, the method of calculating the probability of a POI point being visited based on the inverse distance weighted method specifically includes:

[0028]

[0029] Where sk For S ij The kth POI point in Point s for POI k The probability of the location being visited, k = 0, 1, 2, ..., K, K is S ij The total number of POI points within the area.

[0030] Based on the above technical solution, preferably, the step of assigning time weights to each POI type specifically includes:

[0031] For POI type classification, it is very likely that different POI types are visited at different times. A time weight is assigned to each POI type, and the assigned time weight is used as the time probability of the POI point being visited.

[0032] On the basis of the above technical solution, preferably, the location probability and time probability of the POI point being visited are integrated to obtain the probability of the POI point being visited within the potential visiting time and space range as follows:

[0033] The product of the location probability and time probability of a POI point being visited is used to characterize the spatiotemporal uncertainty, and the probability of a POI point being visited within the potential access spatiotemporal range is obtained:

[0034]

[0035] Where, j = i + 1, P k (i, j) is the probability of the kth POI being visited within the potential access time and space range, is the time probability of the kth POI point, is the probability of the kth POI being visited.

[0036] In a second aspect, the present invention discloses a POI visit probability measurement system based on spatiotemporal trajectories, the system comprising:

[0037] Data acquisition module: used to obtain the spatiotemporal trajectory of pedestrians in the target area and generate a spatiotemporal trajectory point sequence;

[0038] POI point screening module: used to construct the potential space-time range of pedestrian visits using the principles of time geography, generate the pedestrian's space-time prism and potential path area, and use the potential path area to screen out POI points within the potential space-time range;

[0039] Probability calculation module: used to calculate the location probability of POI points being visited within the potential access time and space range based on the inverse distance weighting method; assign time weights to each POI type, obtain the time weights corresponding to POI points within the potential access time and space range as the time probability of POI points being visited; and fuse the location probability and time probability of POI points being visited to obtain the probability of POI points being visited within the potential access time and space range.

[0040] A third aspect of the present invention discloses an electronic device comprising: at least one processor, at least one memory, a communication interface, and a bus;

[0041] The processor, memory, and communication interface communicate with each other via the bus.

[0042] The memory stores program instructions that can be executed by the processor, and the processor calls the program instructions to implement the method according to the first aspect of the present invention.

[0043] According to a fourth aspect of the present invention, a computer-readable storage medium is disclosed, wherein the computer-readable storage medium stores computer instructions, and the computer instructions enable a computer to implement the method according to the first aspect of the present invention.

[0044] The present invention has the following beneficial effects compared to the prior art:

[0045] 1) This invention uses the principles of time geography and the space-time prism tool to calculate the uncertain position points between adjacent trajectory anchor points, thereby filling in the omissions of individual space-time trajectories. At the same time, the space-time prism is used as a space-time kernel to explore the potential space-time range of individual mobile objects between known trajectory points through space-time trajectory data and urban POI data. The spatial information and temporal semantic information of POIs are integrated to calculate the space-time probability of POIs being visited, providing a basis for identifying urban hotspots in urban construction, and also providing technical support for finding people and epidemic investigation.

[0046] 2) The present invention decentralizes the time preferences for visiting different types of POIs in different time periods, establishes the time weights of POIs, combines the space-time prism with POIs, and uses a probabilistic method to fuse the spatial information and time information of POIs. The non-spatial time weight is combined as the time probability to correct the location probability, giving the POI access probability spatial and semantic attributes, thereby making the expression of the probability measure more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0048] Figure 1 Schematic diagram of a method for measuring POI visit probability based on spatiotemporal trajectories provided in an embodiment of the present invention;

[0049] Figure 2 A schematic diagram of the spatiotemporal trajectory of pedestrians provided by an embodiment of the present invention;

[0050] Figure 3 A schematic diagram of a pedestrian space-time prism provided by an embodiment of the present invention;

[0051] Figure 4 A schematic diagram of a pedestrian's two-dimensional trajectory and three-dimensional space-time trajectory provided by an embodiment of the present invention, wherein: Figure 4 (a) is a two-dimensional diagram of pedestrian trajectory. Figure 4 (b) is a schematic diagram of the three-dimensional shape of the pedestrian trajectory;

[0052] Figure 5 Schematic diagram of all PPAs and space-time prisms of pedestrians provided by an embodiment of the present invention, wherein: Figure 5 (a) is a schematic diagram of pedestrian PPA. Figure 5 (b) Schematic diagram of the space-time prism of pedestrians;

[0053] Figure 6 Schematic diagram of the PPA and space-time prism of the first pair of anchor points of pedestrians provided in an embodiment of the present invention, wherein: Figure 6 (a) is a schematic diagram of the PPA of the first pair of anchor points (A0, A1) of the pedestrian and the selected POIs. Figure 6 (b) is a schematic diagram of the first space-time prism of a pedestrian;

[0054] Figure 7 The IDW-based pedestrian visit space location probability map and the spatiotemporal uncertainty distribution map of POI points within the potential space-time range of pedestrian visits provided by the embodiment of the present invention are as follows: Figure 7 (a) is a schematic diagram of the probability density distribution of each position in the PPA area of ​​the first pair of anchor points of the pedestrian. Figure 7 (b) Schematic diagram of the spatiotemporal uncertainty measurement of two different types of POIs in the PPA area of ​​the first pair of anchor points (A0, A1) of the pedestrian in the same time period. DETAILED DESCRIPTION

[0055] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0056] The present invention calculates the uncertain spatiotemporal position of pedestrians between known spatiotemporal trajectory points based on the principles of time geography, so as to support the scientific measurement of the spatiotemporal uncertainty of POI points within the potential spatiotemporal range of pedestrian visits. Time geography has two important tools: space-time path and space-time prism. The space-time path uses a straight line to connect two temporally adjacent known trajectory points, expressing the trend of movement; the space-time prism uses these two anchor points as vertices to express the potential space-time range of pedestrians at the maximum possible speed. The prism is the intersection of two inverted cones, and the cone angle is determined by the movement speed. During the two anchor points, the prism contains not only certain position points (anchor points), but also uncertain position points. These position points are all likely to be visited by pedestrians, and therefore need to be considered in a unified manner.

[0057] In urban spaces, people's activities often involve various POIs, and when choosing activities, they often have preferences regarding time. This paper takes a spatiotemporal trajectory approach, encompassing both time and space, and adds temporal semantic information to POIs by adding location probabilities. Spatially, this approach is weighted based on distance, and temporally, based on time periods. The nodes of a spatiotemporal trajectory represent the observed, definite locations of moving objects, while between trajectory nodes, the moving object has unobserved, uncertain locations.

[0058] See also Figure 1 , a POI visit probability measurement method based on spatiotemporal trajectory, the method comprising:

[0059] S1. Obtain the spatiotemporal trajectory of pedestrians in the target area and generate a spatiotemporal trajectory point sequence.

[0060] A space-time path (STP) can be viewed as a time series of geographic spatial locations, where each spatial location can be represented by coordinates or longitude and latitude. Figure 2 The figure shows a schematic diagram of a pedestrian's space-time trajectory. In the three-dimensional space o-xyt, the x-axis and y-axis represent longitude and latitude respectively, and t represents time. The space-time trajectory includes two forms: two-dimensional (dashed line A0A′1A′2A′3) and three-dimensional (solid line A0A1A2A3). The two-dimensional space-time trajectory is the projection of the three-dimensional space-time trajectory on the plane. The pedestrian's trajectory anchor point can be represented as A i (x i ,y i , t i), A i is the pedestrian trajectory point, where (x i ,y i ) is the spatial position of the trajectory point passed by the pedestrian, t i The corresponding spatial location (x i ,y i ) time. The sequence of space-time trajectory points extracted from the space-time trajectory is:

[0061] STP={A0,A1,…,A i , A j ,…,A n} (1)

[0062] Among them, the i-th trajectory point A i The subsequent trajectory point in the tense is recorded as A j , that is, j=i+1, j≤n, i=0, 1, ..., n, and the total number of space-time trajectory points is n+1.

[0063] S2. Use the principles of time geography to construct the potential space-time range of pedestrians, generate the pedestrian's space-time prism and potential path area, and use the potential path area to filter out POI points within the potential space-time range.

[0064] In the principles of time geography, a spacetime prism represents the possible reach of a moving object within the constraints of its starting point, destination, time, maximum speed, and geographic environment. This prism is the set of all possible trajectories for the moving object. Mathematically, the sum of the distances from each spacetime point in the set to its starting point and destination does not exceed the maximum distance a moving object can travel within a specified timeframe.

[0065] The present invention calculates the potential space-time range of pedestrians between two space-time trajectory points of temporal neighbors by using space-time prism technology. i 、A j , and the corresponding times are t i , t j , in [t i , t j ] can be expressed as:

[0066]

[0067] Where z ijt Indicates A i 、A j The space-time prism with anchor points as vertices represents the potential space-time range of pedestrians, that is, in [t i , t j] during the maximum possible access range. p represents a three-dimensional space-time point, i is the space-time trajectory point number, i = 0, 1, 2, ..., n, j = i + 1, j ≤ n, the total number of space-time trajectory points is n + 1, |||·||| represents the Euclidean distance in three-dimensional space, v max is the maximum moving speed of pedestrians;

[0068] Figure 3 Schematic diagram of the space-time prism for pedestrians provided by an embodiment of the present invention. The footprint or projection of the space-time prism in the plane space is used as the potential path area of ​​the pedestrian, and the potential path area is recorded as PPA (potential path area). In the homogeneous space, the potential path area is composed of two trajectory points (A i , A j ) as the focus with v max ·(t i -t j ) is the major axis of an ellipse:

[0069] g ij ={p′||p′-A′ i ||+||A′ j -p′||≤(t i -t j )·v max}

[0070] Where g ij It is prism z ijt The projection in the plane space represents the potential path area of ​​the pedestrian, ||·|| represents the Euclidean distance in the two-dimensional space, p′ represents the position point in the plane space, A′ i , A′ j A i 、A j Projection in plane space.

[0071] According to the calculated pedestrian PPA, we can filter out the POI points within the potential space-time range visited by pedestrians, and prepare for the next step of adding temporal semantic information to POI. Specifically, the set of all POI points in the target area and the potential path area g ij The intersection of , obtains the POI points within the potential space-time range of pedestrians:

[0072] S ij ={s0, s1, ..., s k ,…,s K}=S∩g ij (4)

[0073] Where S ij is the set of POI points within the potential space-time range visited by pedestrians, s kis a single POI point, k = 0, 1, 2, ..., K, K is S ij The total number of POI points in the target area. S is the set of all POI points in the target area.

[0074] When measuring the probability of a POI being visited, spatial factors and temporal factors are equally important, which corresponds to the description of individual behavior in real space. In urban space, individual activities include various types of POI points, and individual activities are carried out at a certain location within a certain time period. The individual's activity trajectory constitutes spatiotemporal trajectory data with time tags, but the spatiotemporal trajectory data only contains known trajectory anchor point information and cannot describe a large amount of unknown location point information between the trajectory anchor points. The time geography spatiotemporal prism tool provides a solution to this problem, calculating the spatiotemporal locations that individuals may potentially visit between known trajectory anchor points. Therefore, the present invention uses the principles of time geography to calculate the potential path area of ​​pedestrians, and uses the potential path area to screen out POI points within the potential visit spatiotemporal range, thereby calculating the POI visit probability.

[0075] S3. Calculate the location probability of POI points being visited within the potential visiting time and space range based on the inverse distance weighted method.

[0076] The space-time prism is an ideal model based on only two variables: the trajectory time budget and the maximum possible speed. Although the space-time prism has a wide range of uses, its main limitation is that when the maximum speed or time budget is large enough, the prism or PPA is too large or even exceeds the study area, thus losing its physical meaning in delineating the accessible domain. In reality, especially when describing human activities, the maximum possible speeds of different travel modes vary greatly, such as walking (5 km / h) and driving (50 km / h, a lower estimate in cities). Only obtaining pedestrian spatiotemporal anchor points without more information about pedestrian movement patterns will result in excessively large maximum possible pedestrian movement speeds. Therefore, it is necessary to expand the prism using probabilistic temporal geography methods.

[0077] Since the probability of motion within the spacetime prism is not uniform, a probability density function can be used to shrink the prism by assigning a probability to each point within the prism to measure the likelihood of the object being in that position at a specific time.

[0078] The present invention calculates the distance between each POI point and two anchor points (A′ i , A′ j ), namely ||xA′ i ||+||A′ j -x||, combined with the inverse distance weighted IDW (Inverse Distance Weighted) to assign the probability of POI points within the PPA, let s k For S ijThe kth POI point in s k Probability of a location being visited for:

[0079]

[0080] Where K is S ij The total number of POI points within.

[0081] S4. Assign a time weight to each POI type, and obtain the time weight corresponding to the POI points within the potential visit time and space range as the time probability of the POI point being visited.

[0082] When people consider activities, they have different preferences for visiting POIs at different times. Based on this preference, we assign a time weight to each POI type. This means that the attractiveness of different POI types varies non-uniformly over time. Therefore, we add time-dependent weights to POIs based on the time period or time tag of the space-time prism.

[0083] For POI type classification, we assign time weights to each POI type based on the preferences of different POI types for being visited at different times. Specifically, referring to the POI classification code of Amap and combining the daily travel preferences of pedestrians, we divide POIs into 10 major categories and add weights for different POI types in different time periods, as shown in Table 1:

[0084] Table 1. Access weights of different types of POIs in different time periods of semantic information.

[0085]

[0086] The present invention obtains the time weight corresponding to the POI point within the potential visit time and space range according to Table 1, and uses the time weight as the time probability of the POI point being visited.

[0087] S5. The location probability and time probability of the POI point being visited are integrated to obtain the probability of the POI point being visited within the potential visiting time and space range.

[0088] The measurement of the spatiotemporal uncertainty of POI points within the potential space-time range of pedestrians should be based on both spatial and temporal perspectives. The spatial uncertainty is PPA, and the temporal uncertainty includes the range of the space-time prism and the preference for different types of POIs in different time periods. Therefore, the spatiotemporal uncertainty can be characterized by the product of the location probability and time probability of the POI point being visited. The calculation formula for the spatiotemporal uncertainty of POI points within the potential space-time range of pedestrians is as follows:

[0089]

[0090] Where, j = i + 1, P k (i, j) is the probability of a POI point being visited within the potential access time and space range, is the time probability of the POI point, is the probability of the POI point being visited.

[0091] The present invention proposes a POI visit probability measurement method based on spatiotemporal trajectory. Through spatiotemporal trajectory data and urban POI data, the potential spatiotemporal range of individual mobile objects between known trajectory points is mined, the spatial information and temporal semantic information of POIs within the potential visit range of individual mobile objects are integrated, and the spatiotemporal probability of POI visits is calculated. This method measures the uncertainty of POI visits based on spatiotemporal trajectory, provides a basis for identifying urban hotspots in urban construction, and also provides technical support for finding people and epidemic investigation.

[0092] This invention has speculative interpretability in its application to finding people. For example, when considering only spatial factors, POIs that are close have a higher probability of being visited. However, people have different preferences for visiting different POIs over different time periods, resulting in different visit probabilities for different POI types, and even higher visit probabilities for POIs that are farther away.

[0093] The existing technology generally counts the frequency of POI visits by analyzing known trajectory data, but no method has been found to calculate the spatiotemporal probability of POI visits within the potential visit range between trajectory points based on spatiotemporal trajectory data.

[0094] Corresponding to the above method embodiment, the present invention further proposes a POI visit probability measurement system based on spatiotemporal trajectory, the system comprising:

[0095] Data acquisition module: used to obtain the spatiotemporal trajectories of pedestrians in the target area and formalize them to generate a sequence of spatiotemporal trajectory points;

[0096] POI point screening module: used to construct the potential space-time range of pedestrian visits using the principles of time geography, generate the pedestrian's space-time prism and potential path area, and use the potential path area to screen out POI points within the potential space-time range;

[0097] Probability calculation module: used to calculate the location probability of POI points being visited within the potential access time and space range based on the inverse distance weighting method; assign time weights to each POI type, obtain the time weights corresponding to POI points within the potential access time and space range as the time probability of POI points being visited; and fuse the location probability and time probability of POI points being visited to obtain the probability of POI points being visited within the potential access time and space range.

[0098] The above system embodiments and method embodiments correspond one to one. For a brief description of the system embodiments, please refer to the method embodiments.

[0099] In order to demonstrate the creativity and technical value of the technical solution of the present invention, the present invention will be further described below in conjunction with application examples on specific products or related technologies.

[0100] Pedestrian trajectories, as spatiotemporal trajectories, are an important basis for scientifically measuring the spatiotemporal uncertainty of pedestrians' potential visits to POIs within their potential spatiotemporal range. Their accuracy directly impacts the measurement results. Spatiotemporal trajectories can be viewed as a time series of geographic locations, with each location represented by coordinates or latitude and longitude. Figure 3 The space-time prism represented is, to a certain extent, a space-time kernel, which uses a prism to express the influence range of a line segment. Similar to the kernel density function in kernel density estimation, the kernel is a characterization of the uncertainty influence of a known point or line on the surrounding area. Space-time kernels are generally divided into point kernels and line kernels. Among them, the point kernel only considers the influence range of the point and ignores the influence of the line between the anchor points. The line kernel is essentially a buffer body, but the buffer circle is different. Between the two anchor points, the prism contains not only the determined position points (anchor points) but also the uncertain position points. These position points may be visited by pedestrians and therefore need to be considered together.

[0101] The space-time path in time geography can well represent the space-time trajectory of pedestrians. By extracting each pair of activity positions and time anchor points, these positions and time anchor points are strictly sorted in time order. Combined with time geography, according to formula (1), the movement trajectory of pedestrians can be simulated in three-dimensional space, that is, the space-time trajectory of pedestrians.

[0102] The space-time prism represents the area that a mobile object may reach under the common constraints of the starting point, end point, time, maximum speed, and geographical environment. It is the set of all possible movement trajectories of the mobile object. Mathematically, the sum of the distances of each space-time position point element in the set to the starting point and end point does not exceed the maximum movement distance of the mobile object within the limited time. After extracting the pedestrian's space-time trajectory, the pedestrian's space-time prism can be constructed according to formula (2) to characterize the space-time range that pedestrians can potentially visit at the anchor point and between each pair of anchor points. The projection of the space-time prism on the two-dimensional plane is the range expressed by formula (3), namely PPA. According to formula (4), the intersection of PPA and all POIs within the study range can be used to screen out the potential POI points that pedestrians can visit between each pair of anchor points.

[0103] In the real world, the space-time prism is an ideal model, especially when describing human activities. The maximum possible speeds of different travel modes vary greatly. Simply obtaining a pedestrian's space-time anchor point without more information about their movement patterns can oversize a pedestrian's potential access range, which is inconsistent with common sense. Therefore, it is necessary to expand the prism using probabilistic temporal geography methods. This paper uses inverse distance weighting (IDW) to assign probabilities to PPAs. Based on formula (6), the probability density of each location within the PPA can be calculated. The spatial weight of a POI is equal to the probability density of its location.

[0104] In urban spaces, people's activities often encompass various POIs, and when choosing activities, they often have time preferences. Based on this preference, we assign a time weight to each POI type. This means that the attractiveness of different POI types varies non-uniformly over time. Therefore, we assign time-dependent weights to POIs based on the time periods or time tags in the space-time prism.

[0105] The measurement of the spatiotemporal uncertainty of POI points within the potential space-time range of pedestrians should be based on both spatial and temporal perspectives. The spatial uncertainty is PPA, and the temporal uncertainty includes the range of the space-time prism and the preference for different types of POIs in different time periods. Therefore, the product of the two can be used to characterize the spatiotemporal uncertainty of POI visits.

[0106] The present invention has achieved some positive results during the development or use process and does have great advantages compared with the existing technology. The following content describes it with reference to the data, charts, etc. of the experimental process.

[0107] As a preferred embodiment, the method for measuring POI visit uncertainty provided by the embodiment of the present invention specifically includes the following steps:

[0108] Step 1: Obtain the spatiotemporal trajectory of pedestrians.

[0109] Obtain the spatiotemporal trajectory of the pedestrian and formalize it according to formula (1) to generate a sequence of spatiotemporal trajectory points.

[0110] The data for this case comes from the official website of a local health commission (http: / / wjw.enshi.gov.cn / ) and the official website of NetEase (https: / / www.163.com / dy / article / HLAVPO680534B917.html). It primarily consists of a pedestrian trajectory, and the data format is shown in Table 2. Using a Python crawler, we retrieved 82,913 POI data from a local map, including attributes such as name, major category, medium category, province, city, county, and geographic coordinates.

[0111] Table 2 Pedestrian spatiotemporal trajectory data examples

[0112] serial number longitude latitude time address 0 109.4893 30.2678 2022-10-268:10 XXX Community 1 109.5016 30.2743 2022-10-268:18 XXXXX Wholesale Market-Building 20 2 109.5044 30.2825 2022-10-268:52 XX Logistics Park 3 109.4725 30.2700 2022-10-269:25 XXX Supermarket (XXX Store) 4 109.5021 30.3080 2022-10-2610:06 XXX Supermarket (XXX Store) 5 109.4909 30.3038 2022-10-2610:42 XXX Supermarket (XXX Store) 6 109.4909 30.3038 2022-10-2610:59 XXX Supermarket (XXX Store) 7 109.5044 30.2825 2022-10-2611:44 XX Logistics Park 8 109.5016 30.2743 2022-10-2612:03 XXXXX Wholesale Market-Building 20

[0113] Figure 4 (a) shows the two-dimensional morphology of pedestrian trajectories. Figure 4 (b) is a schematic diagram of the three-dimensional shape of the pedestrian trajectory.

[0114] Step 2: Use the principles of time geography to calculate the space-time prism and PPA, and filter the POIs in the PPA.

[0115] First, for two temporal neighbor space-time trajectory point pairs (A i , A j ), based on the pedestrian's maximum moving speed v max =1.5m / s, and use the time geography principle of formula (2) to construct the pedestrian’s i , t j ]'s maximum possible access range z ijt The potential path area g of the pedestrian space-time prism in the plane space is generated by formula (3): ij . Figure 5 (a) shows the schematic diagram of pedestrian PPA, Figure 5 (b) Schematic diagram of the space-time prism of pedestrians.

[0116] Then, based on all POI points and potential path areas g in the target area ij , according to formula (4), the POI point set S within the potential space-time range of pedestrian visits is screened out. Figure 6 (a) shows the PPA of the first pair of anchor points (A0, A1) of the pedestrian and the schematic diagram of the filtered POI. Figure 6 (b) Schematic diagram of the first space-time prism of pedestrians.

[0117] Step 3: Calculate the location probability of the POI being visited.

[0118] According to each POI point s k To two anchor points (A i , A j ), namely ||s k -A i ||+||A j -s k ||, combined with the inverse distance weight IDW, the probability of assigning POI points within the PPA is calculated according to formula (5). Figure 7 The IDW-based pedestrian visit space location probability map and the spatiotemporal uncertainty distribution map of POI points within the potential space-time range of pedestrian visits provided by the embodiment of the present invention are as follows: Figure 7(a) Schematic diagram showing the probability density distribution of each position within the PPA area of ​​the first pair of anchor points of the pedestrian.

[0119] Step 4: Calculate the time probability of a POI being visited.

[0120] Based on the POI classification codes in Amap and the preferences of pedestrians in daily travel, POIs were divided into 10 categories. Based on people's preferences for different POI types at different time periods, time weights were added to different POI types at different time periods. The time weights of the selected POIs were used as the time probability of the corresponding POI being visited.

[0121] Step 5: Calculate the spatiotemporal probability of a POI being visited.

[0122] The measurement of the spatiotemporal uncertainty of POI points within the potential space-time range of pedestrians should be based on both spatial and temporal perspectives. The spatial uncertainty is PPA, and the temporal uncertainty includes the range of the spatiotemporal prism and the preference for different types of POIs in different time periods. Therefore, the product of the two can be used to characterize the spatiotemporal uncertainty. The spatiotemporal uncertainty of POI points within the potential space-time range of pedestrians is calculated according to formula (6) as PPA. k (i, j). Figure 7 (b) shows a schematic diagram of the spatiotemporal uncertainty measurement of two different types of POIs within the PPA area of ​​the pedestrian's first pair of anchor points (A0, A1) during the same time period. The two POI types are respectively catering and daily life services. Within the potential visit time and space range of the pedestrian's first pair of anchor points, the probability density value of daily life services is greater than that of catering services. However, during the time period of 11:00-13:00, people prefer catering, so the time distribution of catering is greater than that of daily life services. The product of the two is used to measure the spatiotemporal uncertainty of POI visits within the pedestrian's potential visit time and space range. The results show that although the probability density value of daily life services is greater than that of catering services, pedestrians are more likely to appear at the catering location during this time period.

[0123] The present invention also discloses an electronic device, comprising: at least one processor, at least one memory, a communication interface and a bus; wherein the processor, memory and communication interface communicate with each other through the bus; the memory stores program instructions that can be executed by the processor, and the processor calls the program instructions to implement the aforementioned method of the present invention.

[0124] The present invention also discloses a computer-readable storage medium storing computer instructions that cause the computer to implement all or part of the steps of the method described in the embodiments of the present invention. The storage medium includes various media capable of storing program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0125] The system embodiment described above is merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be distributed across multiple network units. A person skilled in the art may, without inventive effort, select some or all of the modules as needed to achieve the objectives of this embodiment.

[0126] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A POI visit probability measurement method based on spatiotemporal trajectory, characterized by: The method comprises: Obtain the spatiotemporal trajectory of pedestrians in the target area and generate a spatiotemporal trajectory point sequence; Using the principles of time geography to construct the potential space-time range of pedestrian visits, generate the pedestrian's space-time prism and potential path area, and use the potential path area to filter out POI points within the potential space-time range. Using the principles of time geography to construct the potential space-time range of pedestrian visits, generate the pedestrian space-time prism and potential path area specifically includes: No. Track points The subsequent trajectory points in the tense are recorded as ,Right now 、 are two space-time trajectory pairs of temporal neighbors, 、 The corresponding times are ; Based on the pedestrian's maximum movement speed, the principle of time geography is used to construct the potential space-time range of pedestrians' visits within two space-time trajectory points of temporal neighbors: ; Where, Indicates A space-time prism with anchor points as vertices, representing the potential space-time range of pedestrian visits; p represents a three-dimensional space-time point, i Space-time trajectory point number, i =0,1,2,..., n , j = i +1, j≤n , the total number of space-time trajectory points is n+ 1 ， represents the Euclidean distance in three-dimensional space, is the maximum moving speed of pedestrians; The footprint or projection of the space-time prism in the plane space serves as the potential path area for pedestrians: ; Where, It's a prism The projection in the plane space represents the potential path area of ​​the pedestrian. represents the Euclidean distance in two-dimensional space, p' Represents a point in plane space, 、 They are Projection in plane space; The inverse distance weighted method is used to calculate the probability of POI points being visited within the potential space-time range. Assign a time weight to each POI type, and obtain the time weight corresponding to the POI points within the potential access time and space range as the time probability of the POI point being visited; The location probability and time probability of POI points being visited are integrated to obtain the probability of POI points being visited within the potential visiting time and space range.

2. The POI visit probability measurement method based on spatiotemporal trajectory according to claim 1 is characterized in that: The spatiotemporal trajectory is a time series of spatial positions. In the spatiotemporal trajectory point sequence, each spatiotemporal trajectory point includes the spatial position and corresponding time of a trajectory point passed by the pedestrian.

3. The POI visit probability measurement method based on spatiotemporal trajectory according to claim 1 is characterized in that: The method of using the potential path area to filter out POI points within the potential space-time range of pedestrians specifically includes: The collection of all POI points in the target area and the potential path area The intersection of , obtains the POI points within the potential space-time range of pedestrian visits: ; Where, is the set of POI points within the potential space-time range visited by pedestrians, It is the set of all POI points in the target area.

4. The POI visit probability measurement method based on spatiotemporal trajectory according to claim 3 is characterized in that: The method of calculating the probability of a POI point being visited based on the inverse distance weighted method specifically includes: ; Where, for The first k POI points, POI The probability of a location being visited, k =0,1,2,..., K , for The total number of POI points within the area.

5. The POI visit probability measurement method based on spatiotemporal trajectory according to claim 1 is characterized in that: The method of assigning time weight to each POI type specifically includes: For POI type classification, it is very likely that different POI types are visited at different times. A time weight is assigned to each POI type, and the assigned time weight is used as the time probability of the POI point being visited.

6. The POI visit probability measurement method based on spatiotemporal trajectory according to claim 4 is characterized in that: The probability of POI points being visited within the potential space-time range is obtained by integrating the location probability and time probability of POI points being visited: The product of the location probability and time probability of a POI point being visited is used to characterize the spatiotemporal uncertainty, and the probability of a POI point being visited within the potential access spatiotemporal range is obtained: ; Where, j=i +1, The first k The probability of a POI being visited, For the k The time probability of a POI point, For the k The probability of a POI point being visited.

7. A POI visit probability measurement system based on spatiotemporal trajectory, characterized by: The system comprises: Data acquisition module: used to obtain the spatiotemporal trajectory of pedestrians in the target area and generate a spatiotemporal trajectory point sequence; POI point screening module: used to construct the potential space-time range of pedestrian visits using the principles of time geography, generate the pedestrian's space-time prism and potential path area, and use the potential path area to filter out POI points within the potential space-time range. The construction of the potential space-time range of pedestrian visits using the principles of time geography and the generation of the pedestrian space-time prism and potential path area specifically include: No. Track points The subsequent trajectory points in the tense are recorded as ,Right now 、 are two space-time trajectory pairs of temporal neighbors, 、 The corresponding times are ; Based on the pedestrian's maximum movement speed, the principle of time geography is used to construct the potential space-time range of pedestrians' visits within two space-time trajectory points of temporal neighbors: ; Where, Indicates A space-time prism with anchor points as vertices, representing the potential space-time range of pedestrian visits; p represents a three-dimensional space-time point, i Space-time trajectory point number, i =0,1,2,..., n , j = i +1, j≤n , the total number of space-time trajectory points is n+ 1 ， represents the Euclidean distance in three-dimensional space, is the maximum moving speed of pedestrians; The footprint or projection of the space-time prism in the plane space serves as the potential path area for pedestrians: ; Where, It's a prism The projection in the plane space represents the potential path area of ​​the pedestrian. represents the Euclidean distance in two-dimensional space, p' Represents a point in plane space, 、 They are Projection in plane space; Probability calculation module: used to calculate the location probability of POI points being visited within the potential access time and space range based on the inverse distance weighting method; assign time weights to each POI type, obtain the time weights corresponding to POI points within the potential access time and space range as the time probability of POI points being visited; and fuse the location probability and time probability of POI points being visited to obtain the probability of POI points being visited within the potential access time and space range.

8. An electronic device, characterized in that: include: at least one processor, at least one memory, a communication interface, and a bus; The processor, memory, and communication interface communicate with each other via the bus. The memory stores program instructions that can be executed by the processor, and the processor calls the program instructions to implement the method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer instructions, and the computer instructions enable a computer to implement the method according to any one of claims 1 to 6.

Images