CN111666971A - Event identification method, device and equipment based on position location and storage medium - Google Patents
Event identification method, device and equipment based on position location and storage medium Download PDFInfo
- Publication number
- CN111666971A CN111666971A CN202010345878.4A CN202010345878A CN111666971A CN 111666971 A CN111666971 A CN 111666971A CN 202010345878 A CN202010345878 A CN 202010345878A CN 111666971 A CN111666971 A CN 111666971A
- Authority
- CN
- China
- Prior art keywords
- event
- grid
- map
- identified
- historical
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/29—Geographical information databases
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/18—Network design, e.g. design based on topological or interconnect aspects of utility systems, piping, heating ventilation air conditioning [HVAC] or cabling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Geometry (AREA)
- Databases & Information Systems (AREA)
- Computer Networks & Wireless Communication (AREA)
- Computational Mathematics (AREA)
- Multimedia (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Computer Hardware Design (AREA)
- Remote Sensing (AREA)
- Image Analysis (AREA)
Abstract
The application relates to an artificial intelligence technology, belongs to the field of safety protection, and discloses an event identification method, device, equipment and storage medium based on position location, wherein the method comprises the following steps: acquiring a preset map and event information of an event to be identified, wherein the event information comprises position information, and the preset map comprises a plurality of map grids; determining a target map grid where the event to be identified is located according to the position information based on the preset map, and acquiring an identification model corresponding to the target map grid; performing event recognition on the event to be recognized by using the recognition model so as to output an abnormal score corresponding to the event to be recognized; and if the abnormal score is higher than a preset threshold value, determining that the event to be identified is an abnormal event, and improving the accuracy and efficiency of abnormal event identification.
Description
Technical Field
The present application relates to the field of anomaly identification, and in particular, to a method, an apparatus, a device, and a storage medium for identifying an event based on location positioning.
Background
At present, most of abnormal events are identified by manual verification, namely, abnormal events are identified manually according to experience judgment, but the abnormal events are identified manually according to historical experience and are influenced by investigation conditions and artificial subjective factors, so that the identification efficiency is low, and the identification accuracy is not high.
Therefore, how to improve the accuracy and efficiency of abnormal event identification becomes an urgent problem to be solved.
Disclosure of Invention
The application provides an event identification method, an event identification device, event identification equipment and a storage medium based on position location, so as to improve the accuracy and efficiency of abnormal event identification.
In a first aspect, the present application provides a method for event identification based on position location, where the method includes:
acquiring a preset map and event information of an event to be identified, wherein the event information comprises position information, and the preset map comprises a plurality of map grids;
determining a target map grid where the event to be identified is located according to the position information based on the preset map, and acquiring an identification model corresponding to the target map grid;
performing event recognition on the event to be recognized by using the recognition model so as to output an abnormal score corresponding to the event to be recognized;
and if the abnormal score is higher than a preset threshold value, determining that the event to be identified is an abnormal event.
In a second aspect, the present application further provides a device for identifying an event based on position location, the device comprising:
the information acquisition module is used for acquiring a preset map and event information of an event to be identified, wherein the event information comprises position information, and the preset map comprises a plurality of map grids;
the grid determining module is used for determining a target map grid where the event to be identified is located according to the position information based on the preset map and acquiring an identification model corresponding to the target map grid;
the anomaly identification module is used for carrying out event identification on the event to be identified by utilizing the identification model so as to output an anomaly score corresponding to the event to be identified;
and the abnormality determining module is used for determining the event to be identified as an abnormal event if the abnormality score is higher than a preset threshold value.
In a third aspect, the present application further provides a computer device comprising a memory and a processor; the memory is used for storing a computer program; the processor is configured to execute the computer program and to implement the event identification method based on position location as described above when the computer program is executed.
In a fourth aspect, the present application further provides a computer-readable storage medium, which stores a computer program, which when executed by a processor causes the processor to implement the location-based event recognition method as described above.
The application discloses an event identification method, device, equipment and storage medium based on position location, which comprises the steps of obtaining a preset map and event information of an event to be identified, wherein the event information comprises position information, then determining a target map grid where the event to be identified is located according to the position information, obtaining an identification model corresponding to the target map grid, then carrying out event identification on the event to be identified by utilizing the identification model so as to output an abnormal score corresponding to the event to be identified, and finally determining the event to be identified as an abnormal event if the abnormal score is higher than a preset threshold value. Because most normal events in the same grid have common characteristics, the events are divided according to grid latitudes, and the identification model based on each grid is used for carrying out abnormal identification on the event to be identified in the grid, so that the accuracy and the identification efficiency of the abnormal identification on the event are improved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
FIG. 1 is a schematic flow chart diagram of a training method for recognition models provided by an embodiment of the present application;
FIG. 2 is a schematic flow diagram of sub-steps of a training method of the recognition model provided in FIG. 1;
FIG. 3 is a schematic flow chart diagram of a training method of a mesh model provided in an embodiment of the present application;
fig. 4 is a schematic flowchart of an event identification method based on position location according to an embodiment of the present application;
FIG. 5 is a schematic flow chart diagram of another method for identifying an event based on position location according to an embodiment of the present application;
FIG. 6 is a schematic block diagram of a training apparatus that provides a recognition model according to an embodiment of the present application;
FIG. 7 is a schematic block diagram of a training apparatus for a mesh model provided by an embodiment of the present application;
FIG. 8 is a schematic block diagram of an event recognition apparatus based on position location according to an embodiment of the present application;
fig. 9 is a schematic block diagram of a computer device according to an embodiment of the present application.
Detailed Description
The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are some, but not all, embodiments of the present application. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
The flow diagrams depicted in the figures are merely illustrative and do not necessarily include all of the elements and operations/steps, nor do they necessarily have to be performed in the order depicted. For example, some operations/steps may be decomposed, combined or partially combined, so that the actual execution sequence may be changed according to the actual situation.
It is to be understood that the terminology used in the description of the present application herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. As used in the specification of the present application and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.
It should also be understood that the term "and/or" as used in this specification and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.
The embodiment of the application provides an event identification method and device based on position location, computer equipment and a storage medium. The event identification method based on the position location can be used for identifying abnormal events. In the implementation process, the event identification method based on the position location can be applied to traffic management, vehicle parking management, insurance claim payment and the like.
In the present embodiment, for convenience of description, the vehicle insurance fraud event in the insurance claim is taken as an example for detailed description.
Some embodiments of the present application will be described in detail below with reference to the accompanying drawings. The embodiments described below and the features of the embodiments can be combined with each other without conflict.
Referring to fig. 1, fig. 1 is a schematic flow chart of a training method for recognition models according to an embodiment of the present disclosure. According to the training method of the recognition model, the historical events occurring in the map grid are trained, so that the recognition accuracy of the recognition model on the abnormal events occurring in the map grid is improved.
As shown in fig. 1, the training method of the recognition model specifically includes: step S101 and step S102.
S101, respectively acquiring historical event information in each map grid.
Specifically, the historical event information includes historical event features and feature values. For a plurality of map grids in a map, historical event information in each map grid is obtained respectively. The historical event characteristics include, for example, the accident occurrence time of the historical accident occurring in the grid, the accident reason, whether a traffic accident subscription book exists, and the like, and the corresponding characteristic value is the actual value of the characteristic.
S102, training an isolated forest model by using the historical event features and the feature values in each map grid, and taking the isolated forest model obtained by training as an identification model corresponding to each map grid.
Specifically, the technical scheme utilizes an artificial intelligence technology, for each map grid, the historical event characteristics and the characteristic values in the map grid are respectively utilized to train the isolated forest model, and the training process of the isolated forest model can be as follows:
1) randomly selecting (with and without putting back) n sample points (wherein the dimension of the feature is d) from the training set to form m subsets omegai,i∈ 1,2, …, m, and constructing an isolated tree on the m subsets;
2) randomly selecting one feature q and a segmentation value p thereof from d features of n sample points to perform binary segmentation;
3) recursion 2) building an isolated tree until the isolated tree reaches a limit height or only one point in each leaf node;
4) establishing m isolated trees, and defining the abnormal probability of the m isolated trees according to the average height of the m isolated trees;
a) the average path length of each tree is counted:
c(n)＝2H(n-1)-(2(n-1)/n)
b) the probability of defining an anomaly is:
where h (x) is the number of edges passed from the root node to the leaf node of the isolated tree, i.e., the path length. c (n) is h (x) the average of the path lengths at a given number of samples n, which is used to normalize the path length h (x) of sample x. H (k) is a harmonic number, which can be estimated by the formula h (k) ═ ln (k) + ζ, ξ is an euler constant with a value of 0.5772156649, and k is the path length from the root node to the leaf node. E (h (x)) is the expected path length of sample x in a collection of isolated trees.
When E (h (x)) → c (n), s → 0.5, i.e., the average path length of the sample x is close to the average path length of the isolated tree, it is not possible to distinguish whether or not there is an abnormality.
When E (h (x)) → 0, s → 1, that is, the abnormality score of the sample x approaches 1, it is determined to be abnormal.
When E (h (x)) → (n-1), s → 0, it is judged as normal.
In some embodiments, referring to fig. 2, step S102 specifically includes steps S1021 to S1023.
S1021, training the historical event features and the feature values by using a random forest algorithm to obtain the importance of the historical event features; s1022, ranking the importance of the historical event characteristics to adjust the corresponding weight of the historical event characteristics according to a ranking result; and S1023, training an isolated forest model according to the historical event characteristics, the characteristic values and the weights corresponding to the historical event characteristics.
In particular, since the robustness of the isolated forest algorithm is poor when the isolated forest algorithm faces a scene with a large number of features, the frequency of using important features is probably lower than that of some unimportant features, and the result is not accurate enough. Therefore, the historical event features and the corresponding feature values are trained by using a random forest algorithm to obtain the feature importance of all the features. According to the principle of the isolated forest algorithm, the more abnormal point paths are shorter, the more normal point paths are longer, therefore, the importance of the obtained historical event features can be ranked, different weights are distributed to the features according to the ranking result, the higher the importance is, the smaller the distributed coefficient is, when the isolated forest model is trained, the path generated by each feature is multiplied by the coefficient, and therefore in the total path, the isolation function of the features with high importance is larger.
In the training method for the recognition model provided by the embodiment, the historical information in each map grid is obtained, then the isolated forest model is trained by adopting the isolated forest algorithm according to the historical event features and the feature values in the historical information, the isolated forest model obtained through final training is used as the recognition model corresponding to each map grid, and corresponding weight is given to the historical event features when the isolated forest model is trained, so that the recognition accuracy of the recognition model obtained through final training is improved.
Referring to fig. 3, fig. 3 is a schematic flowchart of a training method of a mesh model according to an embodiment of the present application. The method for training the grid model comprises the steps of constructing the grid image of each map grid, and then training the grid model according to the grid image of each map grid, so that the grid model can output the abnormal score of each map grid.
As shown in fig. 3, the training method of the mesh model specifically includes: step S201 and step S202.
S201, obtaining grid information of each map grid, and constructing a grid portrait of each map grid according to the grid information.
Specifically, the grid representation includes grid features and feature values corresponding to the grid features. And acquiring grid information in each map grid, and constructing a grid portrait corresponding to each map grid according to the grid information of each map grid to obtain a plurality of grid portraits.
The grid information comprises the flow of people in a target grid, the proportion of normal cases and fraudulent cases occurring in the grid, the proportion of interest points related to automobiles to the total number of the interest points, the most frequent insurance time period, the average value of the distances between all dangerous places and the reported repair and repair plants in the grid, the proportion of the number of cases of the highest frequent insurance type to the total number of cases and the like.
In a specific implementation process, after the event to be recognized is determined as an abnormal event by using the recognition model, the abnormal event can be further recognized by a scheme such as manual operation, so that a target event is determined from the abnormal event, wherein the target event is the abnormal event which is determined by manual operation and is actually abnormal.
And calculating the proportion of the target event and the abnormal event to calculate the detection accuracy of the recognition model, and constructing the grid portrait according to the grid information after the detection accuracy reaches a threshold value, so that the accuracy of the constructed grid portrait is improved.
S202, respectively training isolated forest models by utilizing the grid features of each map grid and the feature values corresponding to the grid features, and respectively taking the isolated forest models obtained by training as the grid models corresponding to the map grids.
Specifically, the isolated forest model is trained by using the features and the feature values in the grid images of the map grids, and the isolated forest model obtained by training is used as the grid model for outputting the abnormal score of each map grid. After the grid model is obtained, the grid images of each map grid are respectively input into the grid model so that the grid model can output an abnormal score corresponding to each map grid, and the higher the abnormal score is, the more likely an abnormal event occurs in the map grid.
The above embodiment discloses a training method of a grid model, which constructs a grid portrait corresponding to each map grid by obtaining grid information of each map grid, and trains the grid model according to the grid portrait, and can output an abnormal score of each map grid by using the grid model obtained by training, and the higher the abnormal score is, the more likely an abnormal event occurs in the map grid, so as to increase the abnormal recognition strength for the map grid, and when a plurality of events to be recognized occur, the recognition sequence of the plurality of events to be recognized can be determined by using the grid model, and the recognition efficiency of the events to be recognized is improved.
Referring to fig. 4, fig. 4 is a schematic flowchart of an event identification method based on position location according to an embodiment of the present application. According to the event identification method based on the position location, the event to be identified is subjected to the position location through the event to be identified, so that the event to be identified is scored according to the identification model corresponding to the determined target map grid, and the abnormal identification of the event to be identified is realized.
As shown in fig. 4, the method for identifying an event based on position location specifically includes: step S301 to step S304.
S301, obtaining a preset map and event information of an event to be identified, wherein the event information comprises position information, and the preset map comprises a plurality of map grids.
Specifically, the location information refers to an event occurrence location, and the preset map may be a map including a plurality of map grids obtained by grid division of a city map. For a vehicle insurance claim event, the location information refers to the location of the accident. After the accident happens, when the user applies for the car insurance claim, the user can upload the accident happening place by himself. The region, street and the like where the accident happens can be selected during uploading. For example, the accident site is the middle Tibet road in Huangpu district, Shanghai.
In some embodiments, the location information may also be obtained directly by GPS acquisition. In the accident site where the accident occurs, when the user applies for the vehicle insurance claim, the GPS navigation information on the vehicle or the user terminal can be directly acquired, so that the position information of the accident occurrence place is acquired.
In some embodiments, the method for identifying an event based on position location further includes: and carrying out grid division on the map based on the longitude and latitude to obtain a preset map comprising a plurality of map grids.
Specifically, when the map is divided into grids, the map may be divided into a plurality of identical squares, for example, each grid is 100m × 100m in size, and the central point of the divided grid is used as its unique ID, so as to distinguish the plurality of grids.
Wherein the third position after the decimal point of the warp and weft values can be exactly to a hundred meters, i.e. approximately 100 meters offset per 1 change. The entire map is now gridded, each grid is a square, for example, if the coordinates of the four vertices of a grid are (120.123, 32.456), (120.123,32.457), (120.124,32.457), (120.124,32.456), respectively, the center point of the square is taken as its unique ID, i.e., (120.1235, 32.4565).
In some embodiments, when the map is subjected to grid division based on the longitude and the latitude, traffic data such as road traffic and the like can be used as auxiliary information to perform grid division on the map, so that the identification accuracy of the identification model corresponding to the map grid is improved.
S302, based on the preset map, determining a target map grid where the event to be identified is located according to the position information, and acquiring an identification model corresponding to the target map grid.
Specifically, a map grid refers to a plurality of map grids obtained by meshing a map. And each map grid is corresponding to one recognition model, that is, if there are nine map grids, there are nine recognition models, and each map grid corresponds to one recognition model.
After the position information of the event to be identified is obtained, the corresponding longitude and latitude can be called according to the position information, the target map grid where the event to be identified is located is determined based on the longitude and latitude, and after the target map grid is determined, the identification model corresponding to the target map grid is obtained, so that the event to be identified is conveniently subjected to abnormal identification.
In some embodiments, when there are a plurality of events to be identified, the method for identifying an event based on position location further includes:
calling a corresponding grid model based on a target map grid where a plurality of events to be recognized are located, and outputting an abnormal score of each target map grid based on the grid model; and sequencing the events to be identified according to the abnormal scores so as to sequentially identify the abnormal events to be identified.
Specifically, the map grids can be sorted according to the abnormal score of each map grid, when there are multiple events to be identified, the map grid where each event to be identified is located is determined first, and then the events to be identified are subjected to abnormal identification in sequence according to the abnormal score sorting of the map grids. For example, the abnormality recognition may be performed on the event to be recognized occurring in the map grid with the higher ranking of the abnormality scores.
S303, performing event recognition on the event to be recognized by using the recognition model so as to output an abnormal score corresponding to the event to be recognized.
Specifically, the event information of the event to be identified further includes information such as the reason of the event occurrence, the time of the event occurrence, and the like, and taking the vehicle insurance claim event as an example, the event information of the event to be identified includes the time of the event occurrence, the reason of the event, whether a traffic accident subscription book exists, and the like.
And inputting the event information of the event to be identified into the identification model, and performing event abnormal identification on the event to be identified by the identification model, so as to output an abnormal score corresponding to the event to be identified, so as to judge whether the event to be identified is an abnormal event according to the abnormal score.
S304, if the abnormal score is higher than a preset threshold value, determining that the event to be identified is an abnormal event.
Specifically, the preset threshold may be a default value, or may be adjusted according to actual conditions. After the identification model outputs the abnormal score of the event to be identified, if the abnormal score of the event to be identified is larger than a preset threshold value, the event to be identified is determined to be an abnormal event.
In the event identification method based on position location provided in the above embodiment, event information of an event to be identified is obtained, where the event information includes position information, then a target map grid where the event to be identified is located is determined according to the position information, an identification model of the target map grid is obtained, the identification model is used to perform event identification on the event to be identified, so as to output an abnormal score corresponding to the event to be identified, and finally, if the abnormal score is higher than a preset threshold, the event to be identified is determined to be an abnormal event. Because most normal events in the same grid have common characteristics, the events are divided according to grid latitudes, and the identification model based on each grid is used for carrying out abnormal identification on the event to be identified in the grid, so that the accuracy and the identification efficiency of the abnormal identification on the event are improved.
Referring to fig. 5, fig. 5 is a schematic flowchart of another method for identifying an event based on position location according to an embodiment of the present application.
As shown in fig. 5, the method for identifying an event based on position location specifically includes: step S401 to step S405.
S401, acquiring a preset map and event information of an event to be identified, wherein the event information comprises position information and an event image, and the preset map comprises a plurality of map grids.
Specifically, the location information refers to an event occurrence location, and the preset map may be a map including a plurality of map grids obtained by grid division of a city map. The event image refers to a related picture of an event to be identified, such as a historical event or an accident scene picture taken when the event to be identified occurs.
S402, determining a target map grid where the event to be identified is located according to the position information based on the preset map.
Specifically, after the position information of the event to be identified is acquired, the corresponding longitude and latitude can be called according to the position information, and the target map grid where the event to be identified is located is determined based on the longitude and latitude.
And S403, acquiring a historical image of the historical event in the target map grid.
Specifically, the history image refers to a picture related to a history event.
After the target map grid corresponding to the event to be identified is determined, the historical image of the historical event occurring in the target map grid can be obtained, so that image identification can be conveniently carried out according to the historical image.
S404, performing image recognition on the event image to judge the similarity between the event image and the historical image.
Specifically, image recognition is carried out on an event image of an event to be recognized, and the similarity between the event image and a historical image is judged according to a recognition result and the historical image.
In a specific implementation process, the similarity between the event image and the historical image can be calculated in various ways, for example, the similarity between the image and the historical image can be calculated by using a histogram, the similarity between the image and the historical image can be calculated by using a hash value and a hamming distance, the similarity between the image and the historical image can be calculated by using a cosine distance of the image, the similarity between the image and the historical image can be calculated by using a picture structure metric, and the like.
S405, if the similarity between the event image and the historical image is larger than a preset value, determining that the event to be identified is an abnormal event.
Specifically, if the similarity between the event image and the historical image exceeds a preset value, the event image is judged to be similar to the historical image, and the event to be identified is directly determined to be an abnormal event.
And if the similarity between the event image and the historical image is not greater than a preset value, continuing to perform abnormal recognition on the event to be recognized by using the recognition model corresponding to the target map grid.
In the event identification method based on location positioning provided in the above embodiment, the location information and the event image of the event to be identified are acquired, the target map grid where the event to be identified is located is determined according to the location information based on the preset map, then the historical image of the historical event in the target map grid is acquired, and the event image is subjected to image identification to determine the similarity between the event image and the historical image, and if the similarity between the event image and the historical image is greater than the preset value, the event to be identified is determined to be an abnormal event. Identifying the related pictures of the event to be identified, and detecting whether the problem of using the historical pictures exists in a new event occurring in the same grid or not and whether the problem of using the same pictures in different events exists or not. If the problems occur to the event to be identified, the event to be identified is directly determined as an abnormal event, and the calculation amount of the identification model is reduced.
Referring to fig. 6, fig. 6 is a schematic block diagram of a training apparatus for recognition models according to an embodiment of the present application, the training apparatus for recognition models being configured to perform the aforementioned training method for recognition models.
As shown in fig. 6, the training apparatus 500 for identifying a model includes: history acquisition module 501 and model training module 502.
A history obtaining module 501, configured to obtain history event information in each map grid respectively.
And the model training module 502 is configured to train an isolated forest model by using the historical event features and the feature values in each map grid, and use the isolated forest model obtained through training as an identification model corresponding to each map grid.
The model training module 502 includes an importance determination sub-module 5021, a weight adjustment sub-module 5022, and an isolated model sub-module 5023.
Specifically, the importance determination submodule 5021 is configured to train the historical event features and the feature values by using a random forest algorithm to obtain the importance of the historical event features; the weight adjusting submodule 5022 is used for sequencing the importance of the historical event characteristics so as to adjust the weight corresponding to the historical event characteristics according to the sequencing result; and the isolated model submodule 5023 is used for training an isolated forest model according to the historical event characteristics, the characteristic values and the weights corresponding to the historical event characteristics.
Referring to fig. 7, fig. 7 is a schematic block diagram of a training apparatus for a mesh model according to an embodiment of the present application, the training apparatus for the mesh model being configured to perform the aforementioned training method for the mesh model.
As shown in fig. 7, the training apparatus 600 for mesh models includes: a representation construction module 601 and a model training module 602.
The sketch constructing module 601 is used for acquiring grid information of each map grid and constructing a grid sketch of each map grid according to the grid information.
The model training module 602 is configured to train isolated forest models respectively by using the grid features of each map grid and feature values corresponding to the grid features, and use the isolated forest models obtained through training as the grid models corresponding to the map grids respectively.
Referring to fig. 8, fig. 8 is a schematic block diagram of an event recognition device based on position location according to an embodiment of the present application, which is used for executing the aforementioned event recognition method based on position location. Wherein, the event recognition device based on position location can be configured in a server or a terminal.
The server may be an independent server or a server cluster. The terminal can be an electronic device such as a mobile phone, a tablet computer, a notebook computer, a desktop computer, a personal digital assistant and a wearable device.
As shown in fig. 8, the location-based event recognition apparatus 700 includes: an information acquisition module 701, a grid determination module 702, an anomaly identification module 703, and an anomaly determination module 704.
The information obtaining module 701 is configured to obtain a preset map and event information of an event to be identified, where the event information includes location information, and the preset map includes a plurality of map grids.
A grid determining module 702, configured to determine, based on the preset map, a target map grid where the event to be identified is located according to the position information, and obtain an identification model corresponding to the target map grid;
an anomaly identification module 703, configured to perform event identification on the event to be identified by using the identification model, so as to output an anomaly score corresponding to the event to be identified;
an anomaly determination module 704, configured to determine that the event to be identified is an abnormal event if the anomaly score is higher than a preset threshold.
It should be clearly understood by those skilled in the art that, for convenience and brevity of description, the specific working processes of the event identification device based on location positioning and each module described above may refer to corresponding processes in the foregoing embodiment of the event identification method based on location positioning, and are not described herein again.
The above-mentioned event recognition device based on position location can be implemented in the form of a computer program, which can be run on a computer apparatus as shown in fig. 9.
Referring to fig. 9, fig. 9 is a schematic block diagram of a computer device according to an embodiment of the present disclosure. The computer device may be a server or a terminal.
Referring to fig. 9, the computer device includes a processor, a memory, and a network interface connected through a system bus, wherein the memory may include a nonvolatile storage medium and an internal memory.
The non-volatile storage medium may store an operating system and a computer program. The computer program includes program instructions that, when executed, cause a processor to perform any of the location-based event recognition methods.
The processor is used for providing calculation and control capability and supporting the operation of the whole computer equipment.
The internal memory provides an environment for the execution of a computer program on a non-volatile storage medium, which when executed by the processor causes the processor to perform any of the location-based event recognition methods.
The network interface is used for network communication, such as sending assigned tasks and the like. Those skilled in the art will appreciate that the architecture shown in fig. 9 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
It should be understood that the Processor may be a Central Processing Unit (CPU), and the Processor may be other general purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field Programmable Gate Arrays (FPGAs) or other Programmable logic devices, discrete Gate or transistor logic devices, discrete hardware components, etc. Wherein a general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
Wherein, in one embodiment, the processor is configured to execute a computer program stored in the memory to implement the steps of:
acquiring a preset map and event information of an event to be identified, wherein the event information comprises position information, and the preset map comprises a plurality of map grids;
determining a target map grid where the event to be identified is located according to the position information based on the preset map, and acquiring an identification model corresponding to the target map grid;
performing event recognition on the event to be recognized by using the recognition model so as to output an abnormal score corresponding to the event to be recognized;
and if the abnormal score is higher than a preset threshold value, determining that the event to be identified is an abnormal event.
In one embodiment, the processor is further configured to implement:
and carrying out grid division on the map based on the longitude and latitude to obtain a preset map comprising a plurality of map grids.
In one embodiment, the event information further includes an event image, and the processor is further configured to implement:
acquiring a historical image of a historical event in the target map grid;
performing image recognition on the event image to judge the similarity between the event image and the historical image;
and if the similarity between the event image and the historical image is greater than a preset value, determining that the event to be identified is an abnormal event.
In one embodiment, the processor is further configured to implement:
respectively acquiring historical event information in each map grid, wherein the historical event information comprises historical event characteristics and characteristic values;
and training an isolated forest model by using the historical event features and the feature values in each map grid, and using the isolated forest model obtained by training as a corresponding recognition model of each map grid.
In one embodiment, the processor, in implementing the training of an isolated forest model using the historical event features and feature values in each of the map meshes, is configured to implement:
training the historical event features and the feature values by using a random forest algorithm to obtain the importance of the historical event features;
sorting the importance of the historical event characteristics to adjust the weight corresponding to the historical event characteristics according to a sorting result;
and training an isolated forest model according to the historical event characteristics, the characteristic values and the weights corresponding to the historical event characteristics.
In one embodiment, the processor, after implementing the determining of the target map grid where the event to be identified is located according to the location information, is configured to implement:
calling a corresponding grid model based on a target map grid where a plurality of events to be recognized are located, and outputting an abnormal score of each target map grid based on the grid model;
and sequencing the events to be identified according to the abnormal scores so as to sequentially identify the abnormal events to be identified.
In one embodiment, the processor is further configured to implement:
acquiring grid information of each map grid, and constructing a grid portrait of each map grid according to the grid information, wherein the grid portrait comprises grid characteristics and characteristic values corresponding to the grid characteristics;
and respectively training isolated forest models by utilizing the grid features of each map grid and the feature values corresponding to the grid features, and respectively using the isolated forest models obtained by training as the grid models corresponding to the map grids.
The embodiment of the present application further provides a computer-readable storage medium, where a computer program is stored in the computer-readable storage medium, where the computer program includes program instructions, and the processor executes the program instructions to implement any one of the location-based event identification methods provided in the embodiments of the present application.
The computer-readable storage medium may be an internal storage unit of the computer device described in the foregoing embodiment, for example, a hard disk or a memory of the computer device. The computer readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), and the like provided on the computer device.
While the invention has been described with reference to specific embodiments, the scope of the invention is not limited thereto, and those skilled in the art can easily conceive various equivalent modifications or substitutions within the technical scope of the invention. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.
Claims (10)
1. An event identification method based on position location is characterized by comprising the following steps:
acquiring a preset map and event information of an event to be identified, wherein the event information comprises position information, and the preset map comprises a plurality of map grids;
determining a target map grid where the event to be identified is located according to the position information based on the preset map, and acquiring an identification model corresponding to the target map grid;
performing event recognition on the event to be recognized by using the recognition model so as to output an abnormal score corresponding to the event to be recognized;
and if the abnormal score is higher than a preset threshold value, determining that the event to be identified is an abnormal event.
2. The event recognition method of claim 1, further comprising:
and carrying out grid division on the map based on the longitude and latitude to obtain a preset map comprising a plurality of map grids.
3. The event recognition method of claim 1, wherein the event information further comprises an event image, the method further comprising:
acquiring a historical image of a historical event in the target map grid;
performing image recognition on the event image to judge the similarity between the event image and the historical image;
and if the similarity between the event image and the historical image is greater than a preset value, determining that the event to be identified is an abnormal event.
4. The event recognition method of claim 1, further comprising:
respectively acquiring historical event information in each map grid, wherein the historical event information comprises historical event characteristics and characteristic values;
and training an isolated forest model by using the historical event features and the feature values in each map grid, and using the isolated forest model obtained by training as a corresponding recognition model of each map grid.
5. The event recognition method of claim 4, wherein the training of an isolated forest model using the historical event features and feature values in each of the map grids comprises:
training the historical event features and the feature values by using a random forest algorithm to obtain the importance of the historical event features;
sorting the importance of the historical event characteristics to adjust the weight corresponding to the historical event characteristics according to a sorting result;
and training an isolated forest model according to the historical event characteristics, the characteristic values and the weights corresponding to the historical event characteristics.
6. The event recognition method according to claim 1, wherein after determining the target map grid where the event to be recognized is located according to the position information, the method further comprises:
calling a corresponding grid model based on a target map grid where a plurality of events to be recognized are located, and outputting an abnormal score of each target map grid based on the grid model;
and sequencing the events to be identified according to the abnormal scores so as to sequentially identify the abnormal events to be identified.
7. The event recognition method of claim 6, further comprising:
acquiring grid information of each map grid, and constructing a grid portrait of each map grid according to the grid information, wherein the grid portrait comprises grid characteristics and characteristic values corresponding to the grid characteristics;
and respectively training isolated forest models by utilizing the grid features of each map grid and the feature values corresponding to the grid features, and respectively using the isolated forest models obtained by training as the grid models corresponding to the map grids.
8. An event recognition device based on position location, comprising:
the information acquisition module is used for acquiring a preset map and event information of an event to be identified, wherein the event information comprises position information, and the preset map comprises a plurality of map grids;
the grid determining module is used for determining a target map grid where the event to be identified is located according to the position information based on the preset map and acquiring an identification model corresponding to the target map grid;
the anomaly identification module is used for carrying out event identification on the event to be identified by utilizing the identification model so as to output an anomaly score corresponding to the event to be identified;
and the abnormality determining module is used for determining the event to be identified as an abnormal event if the abnormality score is higher than a preset threshold value.
9. A computer device, wherein the computer device comprises a memory and a processor;
the memory is used for storing a computer program;
the processor for executing the computer program and implementing the method for location based event recognition according to any of claims 1 to 7 when executing the computer program.
10. A computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program which, when executed by a processor, causes the processor to implement the method for location-based event recognition according to any one of claims 1 to 7.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010345878.4A CN111666971A (en) | 2020-04-27 | 2020-04-27 | Event identification method, device and equipment based on position location and storage medium |
| PCT/CN2021/077576 WO2021218314A1 (en) | 2020-04-27 | 2021-02-24 | Event identification method and apparatus based on position locating, and device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010345878.4A CN111666971A (en) | 2020-04-27 | 2020-04-27 | Event identification method, device and equipment based on position location and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111666971A true CN111666971A (en) | 2020-09-15 |
Family
ID=72382945
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010345878.4A Pending CN111666971A (en) | 2020-04-27 | 2020-04-27 | Event identification method, device and equipment based on position location and storage medium |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN111666971A (en) |
| WO (1) | WO2021218314A1 (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112289024A (en) * | 2020-10-23 | 2021-01-29 | 杭州海康威视系统技术有限公司 | Method, device, equipment and storage medium for identifying traffic event information |
| CN112883865A (en) * | 2021-02-09 | 2021-06-01 | 北京深蓝长盛科技有限公司 | Ball-bearing breakthrough event identification method and device, computer equipment and storage medium |
| WO2021218314A1 (en) * | 2020-04-27 | 2021-11-04 | 深圳壹账通智能科技有限公司 | Event identification method and apparatus based on position locating, and device and storage medium |
| CN115344658A (en) * | 2022-09-02 | 2022-11-15 | 自然资源部地图技术审查中心 | Method and device for automatically identifying line type problem map |
| CN115618046A (en) * | 2022-09-05 | 2023-01-17 | 中国科学院昆明植物研究所 | Gridding scientific investigation processing method, system, equipment and storage medium |
| CN116132928A (en) * | 2022-12-27 | 2023-05-16 | 曙光云计算集团有限公司 | Event processing method, event processing device, computer equipment, storage medium and product |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114095391B (en) * | 2021-11-12 | 2024-01-12 | 上海斗象信息科技有限公司 | Data detection method, baseline model construction method and electronic equipment |
| CN114170301B (en) * | 2022-02-09 | 2022-05-17 | 城云科技（中国）有限公司 | Abnormal municipal facility positioning method and device and application thereof |
| CN114611616B (en) * | 2022-03-16 | 2023-02-07 | 吕少岚 | Unmanned aerial vehicle intelligent fault detection method and system based on integrated isolated forest |
| CN114580580B (en) * | 2022-05-07 | 2022-08-16 | 深圳索信达数据技术有限公司 | Intelligent operation and maintenance abnormity detection method and device |
| CN115546733B (en) * | 2022-11-23 | 2023-04-18 | 北京数业专攻科技有限公司 | Group behavior characteristic prediction method and device based on mobile signaling |
| CN116304885B (en) * | 2023-05-11 | 2023-08-22 | 之江实验室 | Event identification method, device and equipment based on graph node embedding |
| CN117336210B (en) * | 2023-12-01 | 2024-04-16 | 河北九宸科技有限公司 | Method, device, equipment and storage medium for detecting abnormal flow of internet of things card |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| IT201800005672A1 (en) * | 2018-05-24 | 2019-11-24 | SCALABLE BIOMETRIC ALARM SYSTEM WITH ARTIFICIAL INTELLIGENCE | |
| CN109409430B (en) * | 2018-10-26 | 2021-07-13 | 江苏智通交通科技有限公司 | Traffic accident data intelligent analysis and comprehensive application system |
| CN109410586B (en) * | 2018-12-13 | 2020-06-05 | 中南大学 | Traffic state detection method based on multi-metadata fusion |
| CN110458718A (en) * | 2019-08-09 | 2019-11-15 | 泰康保险集团股份有限公司 | Vehicle insurance cheats recognition methods, device, medium and electronic equipment |
| CN111860561B (en) * | 2019-12-31 | 2024-02-27 | 北京嘀嘀无限科技发展有限公司 | Abnormal stay behavior identification method, device and equipment of network appointment vehicle and storage medium |
| CN111666971A (en) * | 2020-04-27 | 2020-09-15 | 深圳壹账通智能科技有限公司 | Event identification method, device and equipment based on position location and storage medium |
-
2020
- 2020-04-27 CN CN202010345878.4A patent/CN111666971A/en active Pending
-
2021
- 2021-02-24 WO PCT/CN2021/077576 patent/WO2021218314A1/en active Application Filing
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021218314A1 (en) * | 2020-04-27 | 2021-11-04 | 深圳壹账通智能科技有限公司 | Event identification method and apparatus based on position locating, and device and storage medium |
| CN112289024A (en) * | 2020-10-23 | 2021-01-29 | 杭州海康威视系统技术有限公司 | Method, device, equipment and storage medium for identifying traffic event information |
| CN112289024B (en) * | 2020-10-23 | 2022-04-05 | 杭州海康威视系统技术有限公司 | Method, device, equipment and storage medium for identifying traffic event information |
| CN112883865A (en) * | 2021-02-09 | 2021-06-01 | 北京深蓝长盛科技有限公司 | Ball-bearing breakthrough event identification method and device, computer equipment and storage medium |
| CN112883865B (en) * | 2021-02-09 | 2024-01-19 | 北京深蓝长盛科技有限公司 | Identification method and device for break-through event with ball, computer equipment and storage medium |
| CN115344658A (en) * | 2022-09-02 | 2022-11-15 | 自然资源部地图技术审查中心 | Method and device for automatically identifying line type problem map |
| CN115344658B (en) * | 2022-09-02 | 2023-04-28 | 自然资源部地图技术审查中心 | Method and device for automatically identifying line type problem map |
| CN115618046A (en) * | 2022-09-05 | 2023-01-17 | 中国科学院昆明植物研究所 | Gridding scientific investigation processing method, system, equipment and storage medium |
| CN116132928A (en) * | 2022-12-27 | 2023-05-16 | 曙光云计算集团有限公司 | Event processing method, event processing device, computer equipment, storage medium and product |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2021218314A1 (en) | 2021-11-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111666971A (en) | Event identification method, device and equipment based on position location and storage medium | |
| CN111950329A (en) | Target detection and model training method and device, computer equipment and storage medium | |
| CN111581046A (en) | Data anomaly detection method and device, electronic equipment and storage medium | |
| CN112508118B (en) | Target object behavior prediction method aiming at data offset and related equipment thereof | |
| CN109635056B (en) | Power utilization address data processing method and device, computer equipment and storage medium | |
| CN113673530B (en) | Remote sensing image semantic segmentation method, device, computer equipment and storage medium | |
| CN111882877B (en) | Traffic area analysis method and device | |
| CN111178452A (en) | Driving risk identification method, electronic device and readable storage medium | |
| CN109376689B (en) | Crowd analysis method and device | |
| CN112330332B (en) | Methods, computing devices, and media for identifying fraud risk with respect to node tasks | |
| CN114743132A (en) | Target algorithm selection method and device, electronic equipment and storage medium | |
| CN113936082A (en) | Point cloud data rendering method, device, equipment and storage medium | |
| CN112036488A (en) | Event identification method, device and equipment based on image identification and storage medium | |
| CN115840881A (en) | Air data processing method and device and related equipment | |
| CN115830870A (en) | Toll station data processing method and device, computer equipment and storage medium | |
| CN115936762A (en) | Scenic spot pedestrian flow prediction method, device, equipment and storage medium | |
| CN113792671A (en) | Method and device for detecting face synthetic image, electronic equipment and medium | |
| CN112541436B (en) | Concentration analysis method and device, electronic equipment and computer storage medium | |
| CN113486590A (en) | Data processing method, device and storage medium | |
| CN113609990A (en) | Method and device for determining construction progress of target building and server | |
| CN109635057B (en) | Electricity utilization business processing method and device, computer equipment and storage medium | |
| CN109448287B (en) | Safety monitoring method and device, readable storage medium and terminal equipment | |
| CN112597835A (en) | Driving behavior evaluation method and device, electronic equipment and readable storage medium | |
| CN110751567A (en) | Vehicle information processing method, device, computer equipment and storage medium | |
| CN111967579A (en) | Method and apparatus for performing convolution calculation on image using convolution neural network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |