Method, device, electronic device and machine-readable storage medium for identifying a residence point
By obtaining and clustering the location data of construction waste transport vehicles and identifying their residency points, the problem of identifying construction waste violations and locations is solved, the identification efficiency and accuracy are improved, and environmental pollution is reduced.
Patent Information
- Application Number
- CN202111625896.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-12-28
- Publication Date
- 2024-11-26
- Estimated Expiration
- 2041-12-28
AI Technical Summary
It is difficult for the existing technology to efficiently and accurately discover the illegal disposal situation and location of construction waste, resulting in waste of resources and environmental pollution.
By obtaining the location data of the target vehicle, identifying its residency points, and clustering the residency points to determine the final type, it can automatically identify the violation disposal status and location of construction waste.
It improves the efficiency and accuracy of residency point type identification, provides technical support for automatically discovering the disposal of construction waste violations and locations, and reduces resource waste and environmental pollution.
Smart Images

Figure CN114298634B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of data processing technology, and in particular to a method, device, electronic device and machine-readable storage medium for identifying a resident point. Background Art
[0002] Construction waste refers to the abandoned soil, materials and other waste generated during the construction, reconstruction, expansion and demolition of various buildings, structures, pipelines, etc. by construction units and residents' decoration and renovation of houses.
[0003] Illegal disposal of construction waste, such as the indiscriminate dumping of debris, not only causes waste of resources, but also easily creates safety hazards, brings about environmental pollution problems, and endangers the health of nearby people.
[0004] How to efficiently and accurately detect the illegal disposal of construction waste and its location has become a technical problem that needs to be solved urgently. Summary of the invention
[0005] In view of this, the present application provides a residence point identification method, device, electronic device and machine-readable storage medium.
[0006] According to a first aspect of an embodiment of the present application, a method for identifying a residence point is provided, comprising:
[0007] Acquire a station point of each target vehicle in a first time period, wherein the station point is determined based on positioning data of each target vehicle in the first time period;
[0008] Clustering the stationary points of each target vehicle to obtain a plurality of stationary point clusters;
[0009] For any resident point cluster, according to the initial type of each resident point in the resident point cluster, the final type of each resident point in the resident point cluster is determined as the target type.
[0010] According to a second aspect of an embodiment of the present application, a dwell point identification device is provided, comprising:
[0011] An acquisition unit is configured to acquire a stationary point of each target vehicle in a first time period, wherein the stationary point is determined based on positioning data of each target vehicle in the first time period;
[0012] A clustering unit is configured to cluster the stationary points of each target vehicle to obtain a plurality of stationary point clusters;
[0013] The identification unit is configured to determine, for any resident point cluster, a final type of each resident point in the resident point cluster as a target type according to an initial type of each resident point in the resident point cluster.
[0014] According to a third aspect of an embodiment of the present application, there is provided an electronic device, including a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus;
[0015] Memory, used to store computer programs;
[0016] The processor is used to implement the method provided in the first aspect when executing the program stored in the memory.
[0017] According to a fourth aspect of an embodiment of the present application, a computer-readable storage medium is provided, in which a computer program is stored. When the computer program is executed by a processor, the method provided in the first aspect is implemented.
[0018] According to a fifth aspect of an embodiment of the present application, a computer program is provided. The computer program is stored in a computer-readable storage medium, and when a processor executes the computer program, the processor is prompted to execute the method provided in the first aspect.
[0019] The dwelling point identification method of the embodiment of the present application obtains the dwelling points of each target vehicle within the first time period and clusters the dwelling points of each target vehicle to obtain multiple dwelling point clusters. Then, for any dwelling point cluster, based on the initial type of each dwelling point in the dwelling point cluster, the final type of each dwelling point in the dwelling point cluster is determined as the target type, thereby realizing automatic identification of the dwelling point type based on the positioning data, improving the efficiency and accuracy of the dwelling point type identification, and providing technical support for automatically discovering the illegal disposal of construction waste and its location. BRIEF DESCRIPTION OF THE DRAWINGS
[0020] Figure 1 It is a flowchart of a method for identifying a residence point provided in an embodiment of the present application;
[0021] Figure 2 It is a schematic diagram of a dwell point detection process provided in an embodiment of the present application;
[0022] Figure 3 It is a structural schematic diagram of a station point identification device provided in an embodiment of the present application;
[0023] Figure 4 It is a structural schematic diagram of another station point identification device provided in an embodiment of the present application;
[0024] Figure 5 It is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0025] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.
[0026] The terms used in this application are for the purpose of describing specific embodiments only and are not intended to limit this application. The singular forms "a", "said" and "the" used in this application and the appended claims are also intended to include plural forms, unless the context clearly indicates other meanings.
[0027] In order to enable those skilled in the art to better understand the technical solutions provided by the embodiments of the present application and to make the above-mentioned purposes, features and advantages of the embodiments of the present application more obvious and understandable, the technical solutions in the embodiments of the present application are further described in detail below in conjunction with the accompanying drawings.
[0028] It should be noted that the serial numbers of the steps in the embodiments of the present application do not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.
[0029] See also Figure 1 , is a flow chart of a method for identifying a residence point provided in an embodiment of the present application, such as Figure 1 As shown, the residence point identification method may include the following steps:
[0030] Step S100, obtaining a station point of each target vehicle in a first time period; wherein the station point is determined based on positioning data of each target vehicle in the first time period.
[0031] In the embodiment of the present application, the first time period may include one week, half a month, one month or 10 days, etc.
[0032] The target vehicle may include a vehicle used to transport designated items, where the designated items need to be loaded at a registered location and unloaded at a registered location.
[0033] For example, the designated object may be construction waste, which needs to be loaded from a registered construction site and unloaded at a registered disposal site. The target vehicle may be a construction waste transport vehicle, such as a dump truck.
[0034] The positioning data of the target vehicle may be reported by a positioning device installed on the target vehicle, and the positioning data may include but is not limited to GPS (Global Positioning System) data.
[0035] Exemplarily, the positioning data may include, but is not limited to, vehicle identification information, location information (such as latitude and longitude information), and positioning time information.
[0036] In the embodiment of the present application, the positioning data of each target vehicle in the first time period can be obtained, and the stationing point of each target vehicle in the first time period can be determined based on the obtained positioning data of each target vehicle in the first time period.
[0037] Exemplarily, for any target vehicle, the driving trajectory of the target vehicle within the first time period can be determined based on the acquired positioning data of the target vehicle.
[0038] Exemplarily, a piece of positioning data may correspond to a track point in the driving track.
[0039] The station point of the target vehicle can be determined based on the positioning data corresponding to each track point in the driving track of the target vehicle.
[0040] Exemplarily, the dwell point may be identified based on a set time threshold and path threshold (or distance threshold).
[0041] For example, each trajectory point in the driving trajectory of the target vehicle can be traversed, starting from the first trajectory point, and using the trajectory point as an anchor point to calculate the longest sub-trajectory within a distance threshold (which can be called the first distance threshold), and determine whether the time consumption of the sub-trajectory (the time difference between the last trajectory point and the first trajectory point of the sub-trajectory) exceeds the time threshold; if so, the index values of all trajectory points in the sub-trajectory are put into the candidate residence point set as a set, and the next trajectory point after the sub-trajectory is used as a new anchor point, and the above-mentioned method is continued to be processed until all trajectory points are traversed.
[0042] Exemplarily, the position of the candidate dwell point may be an average position of each track point in the sub-track.
[0043] Exemplarily, the dwell time range of the candidate dwell point may be from the earliest positioning time to the latest positioning time of each track point in the sub-track.
[0044] If the time taken for the sub-trajectory does not exceed the time threshold, the next trajectory point of the trajectory point is used as a new anchor point, and the above process is continued until all trajectory points are traversed.
[0045] For example, assuming that the first distance threshold is 200 meters and the time threshold is 300 seconds, when the driving trajectory of the target vehicle in the first time period is determined, the longest sub-trajectory within 200 meters (assuming that it includes trajectory point 1, trajectory point 2, trajectory point 3 and trajectory point 4 in sequence) can be calculated starting from the first trajectory point in the driving trajectory (assuming that it is trajectory point 1), and taking the trajectory point as the anchor point, and determining the time difference between trajectory point 1 and trajectory point 4, if the time difference exceeds the preset 300 seconds, the index values of all trajectory points (trajectory points 1 to 4) in the longest sub-trajectory are calculated. As a set, they are put into the candidate dwelling point set, that is, trajectory points 1 to 4 correspond to one candidate dwelling point, the position of the candidate dwelling point can be the average position of trajectory points 1 to 4, and the next trajectory point after trajectory point 4 (such as trajectory point 5) is used as a new anchor point, and the above processing continues until all trajectory points are traversed; if the time difference between trajectory point 1 and trajectory point 4 does not exceed 300 seconds, the next trajectory point (i.e. trajectory point 2) of the current anchor point (i.e. trajectory point 1) is used as a new anchor point, and the above processing continues until all trajectory points are traversed.
[0046] Among them, for the candidate stay points corresponding to trajectory points 1 to 4, the stay time range of the candidate stay points is from the positioning time corresponding to trajectory point 1 (assuming t1) to the positioning time corresponding to trajectory point 4 (assuming t4), that is, (t1, t4).
[0047] It should be noted that in the embodiment of the present application, when the candidate residence points are obtained in the above manner, the candidate residence points can be merged according to a preset strategy. For example, each candidate residence point can be traversed, and based on the distance between adjacent candidate residence points, adjacent candidate residence points whose distance is within a preset distance threshold (which can be called a second distance threshold) can be merged.
[0048] For the merged adjacent candidate dwelling points, the average position of the two candidate dwelling points is calculated, and the average position is used as the position of the merged candidate dwelling point. The distance determination is continued with the remaining candidate dwelling points to determine whether to merge to obtain the final dwelling point.
[0049] For example, assuming that the candidate stay points include candidate stay points 1 to 4 (in chronological order from earliest to latest), the distance between candidate stay point 1 and candidate stay point 2 can be calculated first. If the distance is less than the second distance threshold (taking 1000 meters as an example), candidate stay point 1 and candidate stay point 2 are merged, and the average position of candidate stay point 1 and candidate stay point 2 is calculated as the position of the merged candidate stay point, and the distance between the merged candidate stay point and stay point 3 is calculated to determine whether the distance is less than 1000 meters, and so on, until all candidate stay points are traversed.
[0050] In addition, in an embodiment of the present application, when the speed of the target vehicle at a certain trajectory point is determined to be 0 based on the positioning data of the target vehicle in the first time period, the trajectory point with a speed of 0 can also be used as the stationary point of the target vehicle in the first time period.
[0051] Step S110: clustering the stationary points of each target vehicle to obtain a plurality of stationary point clusters.
[0052] Step S120: For any resident point cluster, according to the initial type of each resident point in the resident point cluster, determine the final type of each resident point in the resident point cluster as the target type.
[0053] In the embodiment of the present application, considering the actual scenario, when different target vehicles are stationed in the same area, the stationing positions may be different, but the stationing positions of the multiple target vehicles are all associated with the position of the area.
[0054] For example, when a dump truck is loading debris at a construction site, there may be multiple dump trucks loading debris at different locations on the same construction site. At this time, the positions of the stationary points corresponding to the construction site determined based on the positioning data of the multiple dump trucks may be different, but the multiple stationary points all correspond to the same construction site.
[0055] Accordingly, in order to improve the accuracy of residence point type recognition and avoid the final recognition error of the residence point type caused by a single residence point type recognition error, when identifying the type of the residence point, the type recognition result of the single residence point is no longer simply used as the final result. Instead, different residence points can be clustered, and based on the type recognition results of multiple residence points in the same cluster (which can be called the initial type recognition result), the specified residence point type in the cluster (which can be called the target type) is determined as the type of each residence point in the cluster (which can be called the final type).
[0056] Exemplarily, for any resident point cluster, the target type may be an initial type whose statistical characteristics satisfy a specified condition among the initial types of resident points included in the resident point cluster.
[0057] In one example, for any resident point cluster, the initial type with the largest number of resident points in the resident point cluster can be determined as the target type, and the final type of each resident point in the resident point cluster can be determined as the target type.
[0058] Exemplarily, for any resident point cluster, the number of resident points of each initial type can be counted based on the type of each resident point in the resident point cluster (i.e., the initial type), the initial type with the largest number of resident points can be determined as the target type, and the final type of each resident point in the resident point cluster can be determined as the target type.
[0059] In another example, for any resident point cluster, the initial type in which the number of resident points in the resident point cluster exceeds a preset ratio threshold can be determined as the target type, and the final type of each resident point in the resident point cluster can be determined as the target type.
[0060] Exemplarily, for any stay point cluster, the number of stay points of each initial type can be counted based on the initial type of each stay point in the stay point cluster, the initial type whose number of stay points exceeds a preset ratio threshold can be determined as the target type, and the final type of each stay point in the stay point cluster can be determined as the target type.
[0061] In another example, for any residence point cluster, the initial type with the largest number of residence points in the residence point cluster and a proportion exceeding a preset ratio threshold is determined as the target type, and the final type of each residence point in the residence point cluster is determined as the target type.
[0062] Exemplarily, based on the initial types of each resident point in the resident point cluster, the number of resident points of each initial type can be counted, the initial type with the largest proportion of resident points can be determined, and the proportion of resident points of the initial type can be compared with a preset ratio threshold.
[0063] When the proportion of the number of residence points of the initial type (the initial type with the largest proportion of residence points) exceeds the preset ratio threshold, the initial type (referred to as the target type in this article) can be determined as the target type, and the final type of each residence point in the cluster can be determined as the target type.
[0064] It should be noted that in the embodiment of the present application, when the type with the largest number of dwelling points in the cluster is determined, if the number of dwelling points of this type does not exceed the preset ratio threshold, other strategies can be used to further determine the final type of each dwelling point in the cluster. For example, an abnormal prompt can be given so that relevant personnel can make a manual determination.
[0065] It can be seen that in Figure 1 In the method flow shown, the stop points of each target vehicle in the first time period are obtained, and the stop points of each target vehicle are clustered to obtain multiple stop point clusters. Then, for any stop point cluster, based on the initial type of each stop point in the stop point cluster, the final type of each stop point in the stop point cluster is determined as the target type, thereby realizing automatic identification of the stop point type based on the positioning data, improving the efficiency and accuracy of the identification of the stop point type, and providing technical support for the automatic discovery of illegal disposal of construction waste and its location.
[0066] In some embodiments, the dwell point identification solution provided by the embodiments of the present application may also include:
[0067] Dividing the driving trajectory of each target vehicle in the first time period into a plurality of trajectory segments according to the stationing point of each target vehicle in the first time period;
[0068] For any target vehicle, determine the load status of the target vehicle in each trajectory segment;
[0069] According to the load status of each target vehicle in each trajectory segment, the initial type of each residence point is identified respectively.
[0070] Exemplarily, for any target vehicle, when the target vehicle's stopping point is determined, the driving trajectory of the target vehicle in the first time period can be segmented according to the target vehicle's stopping point to obtain multiple trajectory segments (one trajectory segment can be called a trip).
[0071] Considering that the load state of the target vehicle traveling between different types of parking points usually has inherent characteristics.
[0072] Taking a dump truck as an example, when a dump truck is traveling between a construction site and a disposal site, if the dump truck is traveling from the construction site to the disposal site, the dump truck is usually fully loaded; if the dump truck is traveling from the disposal site to the construction site, the dump truck is usually unloaded.
[0073] Therefore, the type (ie, initial type) of each dwelling point can be identified according to the load state of the target vehicle in each trajectory segment, such as an empty state or a fully loaded state.
[0074] It should be noted that in the embodiment of the present application, the fully loaded state and the unloaded state do not limit the weight of the specified items loaded on the target vehicle. The fully loaded state may refer to the state of the target vehicle after loading is completed, and the unloaded state may refer to the state of the target vehicle after unloading is completed.
[0075] Exemplarily, the types of parking points may include a type for the target vehicle to load consignments (referred to herein as the first type), such as a construction site; or a type for the target vehicle to unload consignments (referred to herein as the second type), such as a disposal site.
[0076] In one example, for any target vehicle, determining the load state of the target vehicle in each trajectory segment may include:
[0077] Based on the positioning data of the target vehicle, the features of each trajectory segment of the target vehicle are extracted in a specified dimension, and based on the feature information of each trajectory segment of the target vehicle, a machine learning algorithm is used to determine the load status of the target vehicle in each trajectory segment.
[0078] For example, it is considered that there are usually some differences when the target vehicle is traveling in a fully loaded state or an unloaded state.
[0079] For example, the speed of the target vehicle in a fully loaded state is generally lower than that in an unloaded state; the angular velocity of the target vehicle in a fully loaded state is generally lower than that in an unloaded state (eg, the steering is slower).
[0080] Therefore, in order to improve the accuracy of determining the vehicle's load status, features of specified dimensions can be extracted for each trajectory segment of the target vehicle based on the positioning data corresponding to each trajectory point in the trajectory segment to obtain feature information of each trajectory segment of the target vehicle. Based on the feature information of each trajectory segment of the target vehicle, a machine learning algorithm can be used to determine the load status of the target vehicle in each trajectory segment. Furthermore, the type of residence point can be identified based on the load status of the target vehicle in each trajectory segment in the manner described in the above embodiment.
[0081] In an example, the above-mentioned extracting features of a specified dimension for each trajectory segment of the target vehicle based on the positioning data of the target vehicle may include:
[0082] For any trajectory segment, the dwell time characteristic information of the trajectory segment is determined according to the positioning data corresponding to the first and last dwell points of the trajectory segment.
[0083] For example, it is considered that the time taken to load or unload the consignment by the target vehicle is usually different, and the time taken to load the consignment is usually longer than the time taken to unload the consignment.
[0084] The time it takes for the target vehicle to load and unload the consignment is also related to the length of time the target vehicle stays at different types of stopover points.
[0085] Therefore, in order to accurately determine the load status of the target vehicle in the trajectory segment, for any trajectory segment, the residence time characteristic information of the trajectory segment can be determined based on the positioning data corresponding to the first and last residence points of the trajectory segment, so as to determine the load status of the target vehicle in the corresponding trajectory segment based on the residence time characteristic information of the target vehicle in each trajectory segment.
[0086] Exemplarily, for any dwell point, the dwell time of the dwell point may be determined based on the positioning data corresponding to each track point in the sub-track of the dwell point (the determination method is described above).
[0087] For example, the difference between the earliest timing time and the latest positioning time in the positioning data corresponding to each track point in the sub-track may be determined as the residence time of the residence point.
[0088] For example, still taking the example described in step S100 as an example, for the candidate dwelling points corresponding to trajectory points 1 to 4, assuming that the candidate dwelling point is used as one of the final dwelling points (for example, after the dwelling point merging, it is not merged with other candidate dwelling points), the dwelling time of the dwelling point can be t4-t1.
[0089] Exemplarily, the dwell time characteristic information may include a dwell time identifier, which may be used to characterize the magnitude relationship between the dwell time of the first dwell point and the dwell time of the last dwell point of the trajectory segment.
[0090] For example, considering that the target vehicle's stay time at a stop point for loading consignments (i.e., a first type of stop point) is generally longer than the target vehicle's stay time at a stop point for unloading consignments (i.e., a second type of stop point), therefore, for a trajectory segment, when the stay time at the first stop point of the trajectory segment is longer than the stay time at the last stop point, the probability that the trajectory segment is a target vehicle driving from the first type of stop point to the second type of stop point will be relatively high, that is, the probability that the target vehicle is fully loaded in the trajectory segment will be relatively high.
[0091] Similarly, when the dwell time of the first dwell point of the trajectory segment is shorter than the dwell time of the last dwell point, the probability that the trajectory segment is the target vehicle driving from the second type of dwell point to the first type of dwell point will be relatively high, that is, the probability that the load state of the target vehicle in the trajectory segment is an unloaded state will be relatively high.
[0092] Therefore, the relationship between the dwell time of the first dwell point and the dwell time of the last dwell point of the trajectory segment can be used to predict the load state of the target vehicle in the trajectory segment.
[0093] In another example, performing feature extraction of a specified dimension on each track segment of the target vehicle according to the positioning data of the target vehicle may include:
[0094] For any trajectory segment, the velocity characteristic information and / or angular velocity characteristic information of the target vehicle in the trajectory segment is determined based on the positioning data corresponding to each trajectory point in the trajectory segment.
[0095] For example, it is considered that the target vehicle traveling under different load conditions may have some significant characteristics in terms of speed and angular velocity.
[0096] For example, the speed of the target vehicle in a fully loaded state is generally lower than that in an unloaded state; the angular velocity of the target vehicle in a fully loaded state is generally lower than that in an unloaded state (eg, the steering is slower).
[0097] Therefore, in order to accurately determine the load status of the target vehicle in the trajectory segment, the features of the target vehicle's trajectory segment in the velocity dimension and the angular velocity dimension can be extracted based on the positioning data of the target vehicle.
[0098] Exemplarily, the speed characteristic and angular velocity characteristic information of the target vehicle in the trajectory segment may be determined based on the positioning data corresponding to each trajectory point in the trajectory segment.
[0099] Exemplarily, the acquired positioning data may include not only vehicle identification information, location information and positioning time, but also speed information and direction angle information.
[0100] In one example, the speed feature information includes part or all of the average speed within the trajectory segment, the maximum instantaneous speed, the instantaneous speed variance, the preset maximum value of the average speed per unit time, the preset average speed variance per unit time, and the preset maximum value identifier of the average speed per unit time.
[0101] Exemplarily, the preset maximum average speed per unit time identifier is used to characterize the relationship between the preset maximum average speed per unit time and the target maximum average speed of the trajectory segment, the target maximum average speed is the median of the preset maximum average speed per unit time of each trajectory segment of the vehicle in the second time period, and the first time period includes one or more second time periods.
[0102] For example, considering that there may be significant differences in characteristics such as the average speed, maximum instantaneous speed, and instantaneous speed variance of the target vehicle in an unloaded state or a fully loaded state during driving, the characteristics such as the average speed, maximum instantaneous speed, and instantaneous speed variance of the target vehicle in the trajectory segment can be used to predict the load state of the target vehicle in the trajectory segment.
[0103] In one example, the angular velocity characteristic information includes part or all of the instantaneous angular velocity maximum value, the instantaneous angular velocity variance, and the preset unit time average angular velocity variance within the trajectory segment.
[0104] For example, considering that there may be significant differences in characteristics such as the average angular velocity variance, the instantaneous angular velocity maximum value, and the instantaneous angular velocity variance during the driving process of the target vehicle in an unloaded state or a fully loaded state, therefore, characteristics such as the average angular velocity variance, the instantaneous angular velocity maximum value, and the instantaneous angular velocity variance of the target vehicle in the trajectory segment can be used to predict the load state of the target vehicle in the trajectory segment.
[0105] In some embodiments, the above-mentioned load state may be a predicted probability of a full load state;
[0106] The above method identifies the initial type of each dwelling point based on the load status of each target vehicle in each trajectory segment, including:
[0107] For any stationary point of any target vehicle, when the predicted probability of a full load state of the target vehicle in a trajectory segment preceding the stationary point is greater than a first threshold, and the predicted probability of a full load state of the target vehicle in a trajectory segment following the stationary point is less than a second threshold, the initial type of the stationary point is determined to be the first type;
[0108] When the predicted probability of a full load state of the target vehicle in a previous trajectory segment of the stationary point is less than a second threshold, and the predicted probability of a full load state of the target vehicle in a subsequent trajectory segment of the stationary point is greater than a first threshold, determining that the initial type of the stationary point is the second type;
[0109] The first threshold is greater than the second threshold.
[0110] Exemplarily, the load state of the target vehicle in any trajectory segment can be characterized by the predicted probability of the target vehicle being in a fully loaded state in the trajectory segment.
[0111] For example, a machine learning model for predicting vehicle load status can be constructed, and sample data for training the machine learning model can be constructed, and the constructed machine learning model can be trained using the constructed sample data.
[0112] After completing the training and testing of the machine learning model, the trained learning model can be used to obtain the predicted probability of the load state of the target vehicle in each trajectory segment by taking the feature information of each trajectory segment of the target vehicle extracted in the above manner as input features.
[0113] Exemplarily, for any trajectory segment of any target vehicle, the greater the predicted probability of the load state of the target vehicle in the trajectory segment, the greater the probability that the target vehicle is in a fully loaded state in the trajectory segment; the smaller the predicted probability of the load state of the target vehicle in the trajectory segment, the smaller the probability that the target vehicle is in an unloaded state in the trajectory segment.
[0114] It should be noted that, in the process of constructing the sample data, the characteristic information of each trajectory segment of the target vehicle used for acquiring the sample data may be extracted in the manner described in the above embodiment.
[0115] In addition, the lifting data or load status data of the target vehicle in the same time period can also be obtained to label the sample data, and its specific implementation can be explained in conjunction with specific examples below; alternatively, the sample data can also be labeled by manual labeling.
[0116] Exemplarily, for a parking point for a target vehicle to unload consigned goods (ie, a first type of parking point), when the target vehicle approaches the parking point, it is usually fully loaded; when the target vehicle leaves the parking point, it is usually unloaded.
[0117] For a parking point used for a target vehicle to load consignments (i.e., a second type of parking point), when the target vehicle approaches the parking point, it is usually in an unloaded state; when the target vehicle leaves the parking point, it is usually in a fully loaded state.
[0118] Accordingly, for any stopping point, the type of the stopping point can be determined based on the load status of the target vehicle's previous trajectory segment at the stopping point (i.e., the target vehicle passes through the trajectory segment to drive toward the stopping point) and the load status of the subsequent trajectory segment at the stopping point (i.e., the target vehicle passes through the trajectory segment to leave the stopping point).
[0119] Exemplarily, when the predicted probability of the load state of the target vehicle in the trajectory segment before the stop point is greater than a first threshold value (i.e., the target vehicle is fully loaded), and the predicted probability of the load state of the target vehicle in the trajectory segment after the stop point is less than a second threshold value (i.e., the target vehicle is unloaded), the initial type of the stop point is determined to be the first type.
[0120] When the predicted probability of the load state of the target vehicle in the previous trajectory segment of the staying point is less than the second threshold, and the predicted load state of the target vehicle in the subsequent trajectory segment of the staying point is greater than the first threshold, the initial type of the staying point is determined to be the second type.
[0121] In some embodiments, in step S120, for any resident point cluster, after determining the final type of each resident point in the resident point cluster as the target type according to the initial type of each resident point in the resident point cluster, the following may also be included:
[0122] For any residence point, according to the final type of the residence point and the location information of the residence point, query the registered area information of the type;
[0123] If there is no area information matching the residence point in the registered area information of this type, the residence point is determined to be an unregistered residence point.
[0124] Exemplarily, in order to realize automatic identification of unregistered resident points, when the final type of each resident point is determined in the above manner, the unregistered resident points can be determined based on the final type of each resident point and the area information of each type that has been registered.
[0125] Exemplarily, for any residence point, the registered area information of this type may be queried based on the final type of the residence point and the location information of the residence point.
[0126] For example, assuming that the final type of the stay point is a construction site, the registered construction site information can be queried based on the location information of the stay point to determine whether there is registered construction site information that matches the location information of the stay point.
[0127] Exemplarily, the registered area information may include a location range of the area. For example, the location range of the registered area information may be identified by an electronic fence.
[0128] Exemplarily, it is possible to determine whether the location of the resident point is located within a registered area of this type based on the location information of the resident point and the registered area information of this type. If so, it is determined that the registered area is a registered area that matches the resident point (the information of the registered area is the matching registered area information); otherwise, that is, the location of the resident point is not within any registered area of this type, and it is determined that there is no matching registered area for the resident point.
[0129] Exemplarily, if there is no area information matching the residence point in the registered area information of this type, the residence point is determined to be an unregistered residence point.
[0130] Exemplarily, when unregistered parking points are determined in the above manner, for any target vehicle, when the parking points of the target vehicle include unregistered parking points, the target vehicle is determined to be a target vehicle to be processed, that is, it is determined that the target vehicle may have illegally disposed of designated items (such as construction waste).
[0131] Alternatively, when the number of unregistered parking points included in the parking points of the target vehicle exceeds a preset number threshold, the target vehicle can be determined as a target vehicle to be processed, that is, it is determined that the target vehicle may have illegally disposed of designated items (such as construction waste).
[0132] In order to enable those skilled in the art to better understand the technical solutions provided in the embodiments of the present application, the technical solutions provided in the embodiments of the present application are described below with reference to specific examples.
[0133] In this embodiment, taking the automatic identification scenario of illegal disposal of slag as an example, the target vehicle is a slag truck.
[0134] In this embodiment, the positioning data of the muck truck (taking GPS data as an example) is used to calculate the dwelling point of the muck truck through a dwelling point detection algorithm, the location of the dwelling point and the dwelling time are determined, and the GPS trajectory of each vehicle is divided according to the dwelling point to obtain multiple trajectory segments of each vehicle (one trajectory segment can be called a trip), and features of three dimensions (dwelling point duration dimension, speed dimension and angular velocity dimension) are extracted from each trajectory segment. Model training and prediction are performed to predict the probability of the vehicle being fully loaded or unloaded in each trajectory segment.
[0135] For example, the residence point that changes from an empty state to a fully loaded state can be determined as a construction site, and the residence point that changes from a fully loaded state to an empty state can be determined as a disposal site. Then, based on the regular construction site and disposal site location data (i.e., the registered area information mentioned above), unregistered construction sites and disposal sites (which may be referred to as black construction sites and black disposal sites) are screened out.
[0136] The following is a detailed description of some of the implementation details of the plan.
[0137] In this embodiment, the muck truck's residence point can be identified based on the set time threshold and path threshold (or distance threshold), and the GPS trajectory of the muck truck can be divided into multiple trajectory segments based on the muck truck's residence point. The features of three dimensions (residence point duration dimension, speed dimension, and angular velocity dimension) of each trajectory segment are extracted, and the load status is labeled based on the muck truck's lifting data or load status data to obtain sample data for training a load status prediction machine learning model, and the trained machine learning model is used for muck truck load status prediction.
[0138] 1. Sample construction
[0139] The GPS data of the dump truck within a period of time (i.e., the first time period mentioned above) is obtained. The GPS data may include but is not limited to: the unique identification of the dump truck (vehicle number or license plate number), the longitude of the collection location, the latitude of the collection location, the speed, the direction angle, and the collection time, which is recorded as "data set one".
[0140] The lifting data of the muck truck in the same time period is obtained. The lifting data may include but is not limited to: the unique identification of the muck truck (vehicle number or license plate number), the longitude of the collection location, the latitude of the collection location, the lifting status (the lift_status value may include: 1-flat, 2-lifted) and the collection time, which is recorded as "data set 2", and the vehicle unique identification of "data set 2" is included in "data set 1".
[0141] Exemplarily, the source of vehicle lifting data may include, but is not limited to, reporting vehicle cabin state changes to the on-board host via a sensor device.
[0142] or,
[0143] The load status data of the muck truck in the same time period is obtained. The load status data may include but is not limited to: the unique identification of the muck truck (vehicle number or license plate number), the longitude of the collection location, the latitude of the collection location, the load status (1-heavy load, 0-no load) and the collection time, which is recorded as "data set 2", and the vehicle unique identification of "data set 2" is included in "data set 1".
[0144] Exemplarily, the source of vehicle load status data may include, but is not limited to, identifying whether the vehicle compartment is loaded with cargo through on-board monitoring video, or reporting vehicle weight changes to an on-board host computer through sensor equipment.
[0145] It should be noted that during the sample construction phase, if the vehicle is unable to provide lifting data or load status data, the vehicle load status can also be determined by manually viewing the vehicle monitoring video.
[0146] In this embodiment, by statistically analyzing the GPS data of the muck trucks obtained, it is found that when the muck trucks move back and forth between the construction site and the disposal site, the loading time is about 10-60 minutes and the unloading time is 5-15 minutes. Due to the frequent comings and goings, the moving state and the loading state can be distinguished after the residence point detection. Combined with "Dataset 2", the residence points where lifting has occurred are marked as disposal sites, and the front and rear residence points are marked as construction sites.
[0147] The trajectory segment is cut according to the end time of the previous dwell point to the start time of the next dwell point, which is recorded as "data set three". The trajectory segment from the construction site to the disposal site is judged to be in a full-load state, marked as 1, and the trajectory segment from the disposal site to the construction site is in an unloaded state, marked as 0.
[0148] Exemplarily, the dwell point detection process can be as follows Figure 2 As shown, Figure 2 As shown, the dwell point detection process may include:
[0149] S1, take the first trajectory point of the continuous trajectory as the anchor point.
[0150] Exemplarily, for any vehicle, the continuous trajectory may be a GPS trajectory of the vehicle within a first time period.
[0151] S2. Taking the current anchor point as the starting point, calculate the longest sub-trajectory within the preset distance threshold.
[0152] S3: Determine whether the time taken by the longest sub-trajectory is greater than a preset time threshold. If yes, go to step S4; otherwise, go to step S5.
[0153] S4. The index values of all the trajectory points in the longest sub-trajectory are added as a set to the candidate resident point set, and the next trajectory point after the longest sub-trajectory is determined as a new anchor point.
[0154] Exemplarily, the position of the candidate dwell point corresponding to the longest sub-trajectory can be the average value of the positioning positions corresponding to each trajectory point in the longest sub-trajectory; the time range of the candidate dwell point corresponding to the longest sub-trajectory can be the time range of the positioning time corresponding to each trajectory point in the longest sub-trajectory (from the earliest positioning time to the latest positioning time).
[0155] S5. Determine the next trajectory point of the current anchor point as a new anchor point.
[0156] S6: Determine whether the new anchor point is the last track point of the continuous track. If yes, end the current process; otherwise, go to step S2.
[0157] 2. Feature extraction
[0158] For any trajectory segment of any muck truck, the features of the dwell point duration dimension, speed dimension and angular velocity dimension can be extracted.
[0159] For example, for any track segment of any muck truck, the extracted features may be as shown in Table 1:
[0160]
[0161]
[0162] Among them, the distance calculation of adjacent track points is to calculate the distance between two adjacent GPS points. The formula can be as follows:
[0163]
[0164] Where r is the radius of the earth, lon1 and lat1 are the longitude and latitude of position A, and lon2 and lat2 are the longitude and latitude of position B, respectively.
[0165] Exemplarily, in this embodiment, before constructing samples for model training and testing based on the above-extracted features, the extracted features may be preprocessed.
[0166] Exemplarily, feature preprocessing may include feature outlier processing (such as replacing outliers and null values with 0), normalization, and the like.
[0167] 3. Model training
[0168] The sample data constructed in the above manner is split into a training set and a test set in a certain ratio (such as 8:2) for training and testing the machine learning model.
[0169] Exemplarily, the machine learning model may adopt a random forest classification algorithm.
[0170] For example, after testing, the machine learning model constructed and trained according to the above method has an accuracy rate and a recall rate of about 88% when the number of sample data is 565. That is, the above method can more accurately identify the load status of the dump truck in each track segment.
[0171] 4. Model prediction
[0172] In this embodiment, when a trained machine learning model is obtained in the above manner, real-time GPS data of the dump truck within a certain time period (such as the above-mentioned first time period) in the actual scene can be obtained, and the residence point determination, trajectory segment division, and feature extraction can be performed in the above manner, and the extracted features can be input into the trained machine learning model to obtain the predicted probability of the load state of the dump truck in each trajectory segment.
[0173] Exemplarily, for any trajectory segment of any muck truck, the greater the predicted probability, the greater the probability that the muck truck is fully loaded in the trajectory segment; the smaller the predicted probability, the greater the probability that the muck truck is unloaded in the trajectory segment.
[0174] 5. Model application
[0175] In this embodiment, based on the load status of the muck truck during its travel, a rule is used to determine the initial type of the stay point, and then, based on the initial type of each stay point, the final type of each stay point is determined.
[0176] Exemplarily, the final type determination method of the residence point may include:
[0177] 5.1. Cluster all the residence points within the above time period using a preset clustering algorithm to obtain multiple residence point clusters.
[0178] Exemplarily, the preset clustering algorithm may include but is not limited to a DBSCAN (Density-Based Spatial Clustering of Applications with Noise) algorithm.
[0179] 5.2. Perform rule judgment on the initial type of the residence points in each residence point cluster.
[0180] Exemplarily, the rules may include:
[0181] For a single stay point, if the predicted probability of the previous segment of the journey of the stay point is greater than 0.7 (taking the above-mentioned first threshold as 0.7 as an example), and the predicted probability of the next segment of the journey is less than 0.3 (taking the above-mentioned second threshold as 0.3 as an example), then the initial type of the stay point is determined to be a consumption field.
[0182] If the predicted probability of the previous segment of the stay point is less than 0.3, and the predicted probability of the next segment of the stay point is greater than 0.7, then the initial type of the stay point is determined to be a construction site.
[0183] For any stay point cluster, the proportion of stay points of the consumption field type and / or the proportion of stay points of the construction site type in the stay point cluster are counted.
[0184] When the number of resident points of the consumption field type in the resident point cluster accounts for the largest proportion and exceeds the preset proportion threshold, it is determined that the final type of each resident point in the resident point cluster is a consumption field.
[0185] When the number of stay points of the construction site type in the stay point cluster accounts for the largest proportion and exceeds a preset proportion threshold, it is determined that the final type of each stay point in the stay point cluster is a construction site.
[0186] In this embodiment, when the final type of each residence point is determined in the above manner, the registered area information of the corresponding type can be queried according to the final type of each residence point and the location information of the residence point.
[0187] For example, for a residence point whose final type is a consumption field, the registered consumption field information can be queried based on the location information of the residence point to determine whether the location of the residence point is within the area of any registered consumption field; if so, the residence point is determined to be a registered consumption field; otherwise, that is, the location of the residence point is not within the area of any registered consumption field, and the residence point is determined to be an unregistered consumption field (that is, a black consumption field).
[0188] In the case where unregistered disposal sites and unregistered construction sites are determined in the above manner, the muck trucks whose stationing points include unregistered disposal sites and / or unregistered construction sites can be determined as muck trucks to be processed (i.e., black muck trucks). Alternatively, muck trucks whose number of unregistered disposal sites and unregistered construction sites in the stationing points exceeds a preset number threshold can be determined as muck trucks to be processed.
[0189] It should be noted that in this embodiment, considering the actual scenario, the parking points of the dump trucks include parking lots in addition to disposal sites and construction sites, and the dump trucks may not change from an empty state to a fully loaded state, or from a fully loaded state to an empty state when entering and exiting the parking lot. However, the residence time of the dump trucks in the parking lot is usually longer, and the residence time usually crosses 0 o'clock.
[0190] Correspondingly, for the above-mentioned stay point cluster, when the stay time in the stay point cluster exceeds the preset time threshold (such as 8 hours), and the number of stay points with a stay time spanning 0 o'clock accounts for the largest proportion and exceeds the preset proportion threshold, the final type of each stay point in the stay point cluster can be determined as a parking lot.
[0191] The method provided by the present application is described above. The device provided by the present application is described below:
[0192] See also Figure 3 , is a schematic diagram of the structure of a station point identification device provided in an embodiment of the present application, such as Figure 3 As shown, the station point identification device may include:
[0193] The acquisition unit 310 is configured to acquire a station point of each target vehicle in the first time period, wherein the station point is determined based on the positioning data of each target vehicle in the first time period;
[0194] The clustering unit 320 is configured to cluster the stationary points of each target vehicle to obtain a plurality of stationary point clusters;
[0195] The identification unit 330 is configured to determine, for any resident point cluster, the final type of each resident point in the resident point cluster as the target type according to the initial type of each resident point in the resident point cluster.
[0196] In some embodiments, Figure 4 As shown, the above-mentioned station point identification device may also include:
[0197] The dividing unit 340 is configured to divide the driving trajectory of each target vehicle in the first time period into a plurality of trajectory segments according to the stationing point of each target vehicle in the first time period;
[0198] The determination unit 350 is configured to determine, for any target vehicle, a load state of the target vehicle in each trajectory segment;
[0199] The identification unit 330 is further configured to identify the initial type of each stationary point according to the load status of each target vehicle in each trajectory segment.
[0200] In some embodiments, the determination unit 350 is specifically configured to, for any target vehicle, perform feature extraction of specified dimensions for each trajectory segment of the target vehicle based on the positioning data of the target vehicle, and determine the load status of the target vehicle in each trajectory segment using a machine learning algorithm based on the feature information of each trajectory segment of the target vehicle.
[0201] In some embodiments, the determination unit 350 is specifically configured to determine, for any trajectory segment, the residence time characteristic information of the trajectory segment based on the positioning data corresponding to the first and last residence points of the trajectory segment; the residence time characteristic information includes a residence time identifier, which is used to characterize the relationship between the residence time of the first residence point and the residence time of the last residence point of the trajectory segment.
[0202] In some embodiments, the load state is a predicted probability of a fully loaded state.
[0203] The identification unit 330 is specifically configured to, for any stop point of any target vehicle, determine that the initial type of the stop point is the first type when the predicted probability of the full load state of the target vehicle in a previous trajectory segment of the stop point is greater than a first threshold, and the predicted probability of the full load state of the target vehicle in a subsequent trajectory segment of the stop point is less than a second threshold; when the predicted probability of the full load state of the target vehicle in a previous trajectory segment of the stop point is less than the second threshold, and the predicted full load state of the target vehicle in a subsequent trajectory segment of the stop point is greater than the first threshold, determine that the initial type of the stop point is the second type; wherein the first threshold is greater than the second threshold.
[0204] In some embodiments, the identification unit 330 is further configured to query the registered area information of any residence point based on the final type of the residence point and the location information of the residence point; if there is no area information matching the residence point in the registered area information of the type, the residence point is determined to be an unregistered residence point.
[0205] See also Figure 5 , is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present application. The electronic device may include a processor 501, a communication interface 502, a memory 503 and a communication bus 505. The processor 501, the communication interface 502 and the memory 503 communicate with each other through the communication bus 505. Among them, the memory 503 stores a computer program; the processor 501 can execute the above-described resident point identification method by executing the program stored in the memory 503.
[0206] The memory 503 mentioned herein may be any electronic, magnetic, optical or other physical storage device that may contain or store information, such as executable instructions, data, etc. For example, the memory 503 may be: RAM (Radom Access Memory), volatile memory, non-volatile memory, flash memory, storage drive (such as hard disk drive), solid state drive, any type of storage disk (such as optical disk, DVD, etc.), or similar storage media, or a combination thereof.
[0207] The present application also provides a computer-readable storage medium storing a computer program, such as Figure 5 The computer program may be stored in the memory 503. Figure 5 The processor 501 in the electronic device is executed to implement the above-described residence point identification method.
[0208] The present application also provides a computer program, which is stored in a computer-readable storage medium, such as Figure 5 The computer program is stored in the memory 503, and when the processor executes the computer program, the processor 501 is prompted to perform the residence point identification method described above.
[0209] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.
[0210] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the scope of protection of the present application.
Claims
1. A method for identifying a residence point, characterized in that: include: Acquire a station point of each target vehicle in a first time period, wherein the station point is determined based on positioning data of each target vehicle in the first time period; Clustering the stationary points of each target vehicle to obtain a plurality of stationary point clusters; For any resident point cluster, according to the initial types of each resident point in the resident point cluster, the final type of each resident point in the resident point cluster is determined as the target type; wherein the target type is the initial type of the resident points included in the resident point cluster, the statistical characteristics of which meet the specified conditions; Wherein, the method further comprises: Dividing the driving trajectory of each target vehicle in the first time period into a plurality of trajectory segments according to the stationing points of each target vehicle in the first time period; For any target vehicle, determine the load status of the target vehicle in each trajectory segment; According to the load status of each target vehicle in each trajectory segment, the initial type of each residence point is identified respectively.
2. The method according to claim 1, characterized in that For any target vehicle, determining the load state of the target vehicle in each track segment includes: According to the positioning data of the target vehicle, the features of the specified dimension are extracted for each track segment of the target vehicle, and according to the feature information of each track segment of the target vehicle, the load state of the target vehicle in each track segment is determined by using a machine learning algorithm; Wherein, the load state includes a fully loaded state or an unloaded state; The initial type of each stationary point is identified according to the load state of each target vehicle in each track segment, including: According to whether each target vehicle is in an empty or fully loaded state in each trajectory segment, the initial type of each dwelling point is identified respectively.
3. The method according to claim 2, characterized in that The step of extracting features of a specified dimension from each track segment of the target vehicle based on the positioning data of the target vehicle includes: For any trajectory segment, the residence time characteristic information of the trajectory segment is determined based on the positioning data corresponding to the first and last residence points of the trajectory segment; the residence time characteristic information includes a residence time identifier, and the residence time identifier is used to characterize the relationship between the residence time of the first residence point and the residence time of the last residence point of the trajectory segment.
4. The method according to claim 1, characterized in that The load state is a predicted probability of a fully loaded state; and the initial type of each stationary point is identified according to the load state of each target vehicle in each trajectory segment, including: For any stationary point of any target vehicle, when the predicted probability of a full load state of the target vehicle in a trajectory segment preceding the stationary point is greater than a first threshold, and the predicted probability of a full load state of the target vehicle in a trajectory segment following the stationary point is less than a second threshold, the initial type of the stationary point is determined to be the first type; When the predicted probability of a full load state of the target vehicle in a previous trajectory segment of the stationary point is less than a second threshold, and the predicted probability of a full load state of the target vehicle in a subsequent trajectory segment of the stationary point is greater than a first threshold, determining that the initial type of the stationary point is the second type; Among them, the first threshold is greater than the second threshold, and the predicted probability of the target vehicle in the fully loaded state in the trajectory segment is greater than the first threshold, indicating that the target vehicle is in a fully loaded state, and the predicted probability of the target vehicle in the fully loaded state in the trajectory segment is less than the second threshold, indicating that the target vehicle is in an unloaded state.
5. The method according to any one of claims 1 to 4, characterized in that: For any resident point cluster, after determining the final type of each resident point in the resident point cluster as the target type according to the initial type of each resident point in the resident point cluster, the method further includes: For any residence point, according to the final type of the residence point and the location information of the residence point, query the registered area information of the type; If there is no area information matching the residence point in the registered area information of this type, the residence point is determined to be an unregistered residence point.
6. A station point identification device, characterized in that: include: An acquisition unit is configured to acquire a stationary point of each target vehicle in a first time period, wherein the stationary point is determined based on positioning data of each target vehicle in the first time period; A clustering unit is configured to cluster the stationary points of each target vehicle to obtain a plurality of stationary point clusters; The identification unit is configured to determine, for any resident point cluster, the final type of each resident point in the resident point cluster as a target type according to the initial type of each resident point in the resident point cluster; wherein the target type is an initial type of the resident points included in the resident point cluster whose statistical characteristics meet a specified condition; Wherein, the device further comprises: A dividing unit is configured to divide the driving trajectory of each target vehicle in the first time period into a plurality of trajectory segments according to the stationing points of each target vehicle in the first time period; A determination unit is configured to determine, for any target vehicle, a load state of the target vehicle in each trajectory segment; The identification unit is further configured to identify the initial type of each stay point according to the load status of each target vehicle in each track segment.
7. The device according to claim 6, characterized in that The determination unit is specifically configured to, for any target vehicle, perform feature extraction of a specified dimension on each track segment of the target vehicle according to the positioning data of the target vehicle, and determine the load state of the target vehicle in each track segment using a machine learning algorithm according to the feature information of each track segment of the target vehicle; Wherein, the load state includes a fully loaded state or an unloaded state; The identification unit is specifically configured to identify the initial type of each parking point according to whether each target vehicle is in an empty state or a fully loaded state in each trajectory segment; The determining unit is specifically configured to determine, for any trajectory segment, the dwell time characteristic information of the trajectory segment according to the positioning data corresponding to the first and last dwell points of the trajectory segment; the dwell time characteristic information includes a dwell time identifier, and the dwell time identifier is used to characterize the magnitude relationship between the dwell time of the first dwell point and the dwell time of the last dwell point of the trajectory segment; Wherein, the load state is a predicted probability of a fully loaded state; The identification unit is specifically configured to, for any stationary point of any target vehicle, determine that the initial type of the stationary point is the first type when the predicted probability of the full load state of the target vehicle in a previous trajectory segment of the stationary point is greater than a first threshold, and the predicted probability of the full load state of the target vehicle in a subsequent trajectory segment of the stationary point is less than a second threshold; determine that the initial type of the stationary point is the second type when the predicted probability of the full load state of the target vehicle in a previous trajectory segment of the stationary point is less than a second threshold, and the predicted probability of the full load state of the target vehicle in a subsequent trajectory segment of the stationary point is greater than the first threshold; wherein, the first threshold is greater than the second threshold, the predicted probability of the full load state of the target vehicle in the trajectory segment is greater than the first threshold, indicating that the target vehicle is in a fully loaded state, and the predicted probability of the full load state of the target vehicle in the trajectory segment is less than the second threshold, indicating that the target vehicle is in an unloaded state; and / or, The identification unit is further configured to query the registered area information of any residence point according to the final type of the residence point and the location information of the residence point; if there is no area information matching the residence point in the registered area information of the type, the residence point is determined to be an unregistered residence point.
8. An electronic device, characterized in that: It includes a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other through the communication bus; Memory, used to store computer programs; A processor, for implementing the method according to any one of claims 1 to 5 when executing a program stored in a memory.
9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.
