Graphical contour detection method, device, storage medium, equipment and program product

Abstract

The application discloses a graph contour detection method and device, a storage medium, equipment and a program product, which can be applied to various scenes such as map, location data processing, trajectory data processing and image processing. The method comprises the following steps: obtaining interest surface data and check-in data, each interest surface in the interest surface data containing a corresponding interest surface contour and a geographic location identifier, and the check-in data containing check-in geographic identifiers; calculating a kernel density estimation value corresponding to each interest surface in the interest surface data for a target location point according to a corresponding relationship between the geographic location identifier and the check-in geographic identifier and the interest surface data and the check-in data; generating an isogram of the kernel density estimation value, and generating a plurality of isosurfaces according to the isogram; determining a target isosurface from the plurality of isosurfaces; and detecting an abnormal interest surface contour according to an intersection and union ratio index between a contour of the target isosurface and the interest surface contour, so that preliminary automatic screening can be realized, the labor operation cost of detection is reduced, and the data optimization efficiency is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer, in particular to a graph contour detection method and device, a storage medium, an equipment and a program product. BACKGROUND

[0002] An area of interest (AOI) is an important geographic information element in a map. However, the actual AOI data may have contour abnormalities, such as contour position abnormalities, contour overlarge, contour undersize, etc. Currently, such contour abnormalities are usually classified as abnormal AOI contours. The discrimination of abnormal AOI contours mainly depends on the AOI user, such as active feedback of a map user or manual investigation. It is time-consuming and laborious to manually discriminate abnormal AOI contours, and the cost is huge. SUMMARY

[0003] The embodiments of the present application provide a graph contour detection method and device, a storage medium, an equipment and a program product, which can realize preliminary automatic screening, reduce the manual operation cost of abnormal AOI contour detection, and accelerate the data optimization efficiency.

[0004] In one aspect, a graph contour detection method is provided. The method includes: obtaining AOI data and check-in data, each AOI in the AOI data containing a corresponding AOI contour and a geographic location identifier, the check-in data containing a check-in geographic identifier; calculating a kernel density estimation value of a target location point corresponding to each AOI in the AOI data according to a corresponding relationship between the geographic location identifier and the check-in geographic identifier, and the AOI data and the check-in data; generating an isogram of the kernel density estimation value, and generating a plurality of isosurfaces according to the isogram; determining a target isosurface corresponding to each AOI in the AOI data from the plurality of isosurfaces; and detecting an abnormal AOI contour according to an intersection-union ratio index between the contour of the target isosurface corresponding to each AOI in the AOI data and the AOI contour.

[0005] In another aspect, a graph contour detection device is provided. The device includes:

[0006] An obtaining module is configured to obtain AOI data and check-in data, each AOI in the AOI data containing a corresponding AOI contour and a geographic location identifier, the check-in data containing a check-in geographic identifier.

[0007] A calculating module is configured to calculate a kernel density estimation value of a target location point corresponding to each AOI in the AOI data according to a corresponding relationship between the geographic location identifier and the check-in geographic identifier, and the AOI data and the check-in data.

[0008] generating contours of the kernel density estimation value, and generating a plurality of contour surfaces according to the contours;

[0009] determining a target contour surface corresponding to each interest surface in the interest surface data from the plurality of contour surfaces;

[0010] detecting an abnormal interest surface contour according to an intersection-union ratio index between the contour of the target contour surface corresponding to each interest surface in the interest surface data and the interest surface contour.

[0011] In another aspect, a computer readable storage medium is provided, which stores a computer program adapted to be loaded into a processor to perform the steps in the method for detecting a graphic contour according to any one of the above embodiments.

[0012] In another aspect, a computer device is provided, which comprises a processor and a memory, and the memory stores a computer program, and the processor is configured to perform the steps in the method for detecting a graphic contour according to any one of the above embodiments by invoking the computer program stored in the memory.

[0013] In another aspect, a computer program product is provided, which comprises computer instructions for performing the steps in the method for detecting a graphic contour according to any one of the above embodiments when executed by a processor.

[0014] The embodiments of the present application obtain interest surface data and check-in data, each interest surface in the interest surface data comprises a corresponding interest surface contour and a geographic location identifier, and the check-in data comprises a check-in geographic identifier; calculate a kernel density estimation value corresponding to each interest surface in the interest surface data according to a corresponding relationship between the geographic location identifier and the check-in geographic identifier and the interest surface data and the check-in data; generate contours of the kernel density estimation value, and generate a plurality of contour surfaces according to the contours; determine a target contour surface corresponding to each interest surface in the interest surface data from the plurality of contour surfaces; and detect an abnormal interest surface contour according to an intersection-union ratio index between the contour of the target contour surface corresponding to each interest surface in the interest surface data and the interest surface contour. The embodiments of the present application calculate the kernel density estimation value of each interest surface based on the interest surface data and the check-in data, and calculate the contours and the contour surfaces of the kernel density estimation value, and detect the abnormal interest surface contour according to the intersection-union ratio index between the contour of the determined target contour surface and the interest surface contour, so that the preliminary automatic screening can be realized, the manual operation cost for detecting the abnormal interest surface contour can be reduced, and the data optimization efficiency can be accelerated. BRIEF DESCRIPTION OF DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description only constitute some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained from these drawings without creative effort.

[0016] Figure 1 The structural schematic diagram of the graphic contour detection system provided by the embodiments of the present application is shown.

[0017] Figure 2 The first flowchart of the graphic contour detection method provided by the embodiments of the present application is shown.

[0018] Figure 3 The second flowchart of the graphic contour detection method provided by the embodiments of the present application is shown.

[0019] Figure 4 The third flowchart of the graphic contour detection method provided by the embodiments of the present application is shown.

[0020] Figure 5 The fourth flowchart of the graphic contour detection method provided by the embodiments of the present application is shown.

[0021] Figure 6 The fifth flowchart of the graphic contour detection method provided by the embodiments of the present application is shown.

[0022] Figure 7 The sixth flowchart of the graphic contour detection method provided by the embodiments of the present application is shown.

[0023] Figure 8 The seventh flowchart of the graphic contour detection method provided by the embodiments of the present application is shown.

[0024] Figure 9 The application scenario diagram of the graphic contour detection method provided by the embodiments of the present application is shown.

[0025] Figure 10 The another application scenario diagram of the graphic contour detection method provided by the embodiments of the present application is shown.

[0026] Figure 11 The still another application scenario diagram of the graphic contour detection method provided by the embodiments of the present application is shown.

[0027] Figure 12 The eighth flowchart of the graphic contour detection method provided by the embodiments of the present application is shown.

[0028] Figure 13 The structural schematic diagram of the graphic contour detection device provided by the embodiments of the present application is shown.

[0029] Figure 14 A structural schematic diagram of a computer device provided by an embodiment of the present application is shown. DETAILED DESCRIPTION

[0030] 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. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person skilled in the art without creative work fall within the protection scope of the present application.

[0031] The embodiments of the present application provide a graph contour detection method and device, a storage medium, a computer device and a computer program product. Specifically, the graph contour detection method of the embodiments of the present application can be executed by a computer device, where the computer device can be a terminal or a server and the like. The terminal can be a smart phone, a tablet computer, a notebook computer, a smart voice interaction device, a smart home appliance, a wearable smart device, a smart vehicle terminal, a flying vehicle and the like, and the terminal can further include a client, which can be a video client, a browser client or an instant messaging client and the like. The server can be a standalone physical server, a server cluster or a distributed system composed of multiple physical servers, or a cloud server providing cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content distribution networks (CDN), and basic cloud computing services such as big data and artificial intelligence platforms. The embodiments of the present application can be applied to various scenarios such as the field of maps, location data processing, trajectory data processing, image processing and the like.

[0032] First, some of the nouns or terms appearing in the description of the embodiments of the present application are explained as follows:

[0033] Artificial intelligence (AI) is to use a digital computer or a machine controlled by a digital computer to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use the knowledge to obtain the best results. In other words, artificial intelligence is a comprehensive technology of computer science, which tries to understand the essence of intelligence and produce a new intelligent machine that can react in a similar way to human intelligence. Artificial intelligence is to study the design principles and implementation methods of various intelligent machines, so that the machine has the functions of perception, reasoning and decision-making.

[0034] Computer Vision (CV) Computer vision is a scientific discipline that deals with enabling machines to "see". More specifically, it refers to the use of cameras and computers to identify, inspect, and otherwise "see" objects, and further process the images to make them more suitable for human observation or transmission to instruments for detection. As a scientific discipline, computer vision researches the theories and techniques that attempt to create artificial intelligence systems that can extract information from images or multidimensional data. Computer vision technology typically includes image processing, image recognition, image semantic understanding, image retrieval, OCR, video processing, video semantic understanding, video content / behavior recognition, three-dimensional object reconstruction, 3D technology, virtual reality, augmented reality, simultaneous localization and mapping, and other technologies. It also includes common biometric identification technologies such as face recognition and fingerprint recognition.

[0035] Machine Learning (ML) is a multi-disciplinary subject that involves probability theory, statistics, approximation theory, convex analysis, algorithmic complexity theory, and other disciplines. It is a specialized field that studies how computers can simulate or implement human learning behavior to acquire new knowledge or skills, and reorganize existing knowledge structures to continuously improve their performance. Machine learning is the core of artificial intelligence and the fundamental approach to making computers intelligent. It is applied in various fields of artificial intelligence. Machine learning and deep learning typically include artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and inductive learning.

[0036] Intelligent Traffic System (ITS) is also known as Intelligent Transportation System (ITS). It is a comprehensive transportation system that integrates advanced scientific and technological (information technology, computer technology, data communication technology, sensor technology, electronic control technology, automatic control theory, operations research, artificial intelligence, etc.) into transportation, service control, and vehicle manufacturing, and strengthens the connection between vehicles, roads, and users, to form a comprehensive transportation system that ensures safety, improves efficiency, improves the environment, and saves energy.

[0037] Cloud technology: refers to the series of resources such as hardware, software, network, etc. in the wide area network or local area network are unified, realize the data calculation, storage, processing and sharing of a kind of hosting technology. Cloud technology based on cloud computing business model application of network technology, information technology, integration technology, management platform technology, application technology, etc. can be composed of resource pool, on demand, flexible and convenient. Cloud computing technology will become an important support. The background service of technical network system needs a lot of computing and storage resources, such as video website, picture website and more portal website. With the high development and application of Internet industry, every item may have its own identification mark in the future, which needs to be transmitted to the background system for logical processing. Different levels of data will be processed separately, and various industry data need strong system support, which can only be realized through cloud computing.

[0038] Point of interest (POI, Point Of Interest) is also called information point, which refers to the meaningful point in the map, and the main purpose is to describe the address of things or events. For example, the point of interest can be the location point of a community, a school or a shopping mall.

[0039] Area of interest (AOI, Area of Interest): a regional geographical entity in map data, such as the area of a community, a school, a park, a hospital, etc.

[0040] Location based service (LBS, Location Based Service) is a value-added service that obtains the location information (geographic coordinates or geodetic coordinates) of mobile terminal users through the wireless communication network (such as GSM network, CDMA network) of telecom mobile operators or external positioning method (such as GPS), and provides corresponding services for users under the support of GIS (Geographic Information System) platform.

[0041] LBS check-in: is the behavior of marking the specified geographic coordinates based on LBS, which indicates that the user has visited a place at the current geographic coordinates, such as sending information with the behavior of attaching the place through social circle (such as friends circle, microblog, taoyou circle, etc.).

[0042] Kernel density estimation: Kernel density estimation (KDE) is a density-based non-parametric estimation method widely used in spatial analysis. It assumes that any location within a region has a measurable event density, which can be estimated by the number of event points in its neighborhood. If more events occur in the neighborhood of a location, it is considered that the probability of events occurring at that location is higher. The kernel density estimation method can be compared to using a moving window to calculate the distance of event points within the window to the location, which is considered as the weight of event influence.

[0043] Contour: A smooth curve connecting points with equal values of a quantity of a mapped object. Contour maps are usually drawn by marking the relative point values of a mapped object on a base map, then connecting the points with equal values to form a smooth curve, and outlining the spatial structure characteristics of the mapped object.

[0044] To detect abnormal interest surface profiles, the industry often relies on two methods. One is to rely on feedback from data users, such as mobile map users who report that the interest surface is incorrect, or product users who report that the interest surface is incorrect. This is a post-feedback method that can easily cause data users to lose confidence in the data. The second is to solve the problem through manual operation, but the amount of geographic data such as interest surface is often very large (millions, tens of millions, etc.). Without user feedback, it is difficult to find possible abnormal interest surfaces from the vast amount of interest surface data through limited manpower. Therefore, an effective, reliable, and accurate abnormal interest surface profile detection method is needed to achieve preliminary automatic screening and reduce the cost of manual operation.

[0045] The embodiment of the present application proposes a graph profile detection method based on LBS check-in data and kernel density analysis. First, the LBS check-in data of a certain interest surface can be estimated based on a specified bandwidth. Second, the contour lines and contour surfaces of the kernel density estimation value are drawn, and the most suitable target contour surface is selected according to the IOU index of different contour surfaces and interest surface profiles to obtain the error evaluation result. The above process is iterated by adjusting the bandwidth parameter until the mean of the IOU index of all interest surfaces is minimized, thereby selecting the best bandwidth parameter as the target bandwidth. Then, the IOU index calculated based on the target bandwidth is sorted from small to large to obtain an automatic preliminary screening detection list of abnormal interest surface profiles. The embodiment of the present application can automatically mine potential abnormal interest surface profiles by iteratively selecting the best parameter based on LBS check-in data reflecting human spatial cognition, realize low-cost preliminary screening of abnormal interest surface profiles, effectively reduce the cost of manual operation, reduce the time and human resource investment of abnormal interest surface profile mining and repair, and optimize efficiency.

[0046] Please refer to Figure 1 , Figure 1 The structural schematic diagram of the graphic contour detection system provided by the embodiments of the present application is shown in the figure. The graphic contour detection system includes a terminal 10 and a server 20, etc. The terminal 10 and the server 20 are connected through a network, such as a wired or wireless network connection, etc.

[0047] The terminal 10 can be used to display a graphic user interface. The terminal is used to interact with the user through the graphic user interface, such as downloading and installing a corresponding client through the terminal and running, such as calling and running a corresponding applet, such as logging in to a website to present a corresponding image user interface, etc. In the embodiments of the present application, the terminal 10 can be a terminal device used by the user to upload the original image and display the color matching result. When performing graphic contour detection, the terminal device used by the user can generate check-in data and send the check-in data to the server 20.

[0048] In the embodiments of the present application, when performing graphic contour detection, the server 20 can be specifically used to: obtain interest surface data and check-in data, each interest surface in the interest surface data contains a corresponding interest surface contour and a geographic location identifier, and the check-in data contains a check-in geographic identifier; according to the correspondence between the geographic location identifier and the check-in geographic identifier, and the interest surface data and the check-in data, calculate the kernel density estimation value of the target position point corresponding to each interest surface in the interest surface data; generate an isogram of the kernel density estimation value, and generate a plurality of isosurfaces according to the isogram; determine the target isosurface corresponding to each interest surface in the interest surface data from the plurality of isosurfaces; detect the abnormal interest surface contour according to the intersection and union ratio index between the contour of the target isosurface corresponding to each interest surface in the interest surface data and the interest surface contour; and send the abnormal interest surface contour to the terminal 10 for display.

[0049] In addition, the terminal 10 receives the cause analysis result of the abnormal interest surface contour input by the user according to the displayed abnormal interest surface contour, and sends the cause analysis result to the server 20. The server 20 can also update the interest surface data according to the cause analysis result.

[0050] The following will be described in detail. It should be noted that the order of the following embodiments is not limited as the priority order of the embodiments.

[0051] The embodiments of the present application provide a graphic contour detection method, which can be executed by a terminal or a server, or jointly executed by a terminal and a server. The embodiments of the present application take the graphic contour detection method executed by the server as an example to illustrate.

[0052] Please refer to Figures 2 to 12 ,Figures 2 to 8 , Figure 12 are flowchart diagrams of a graphic contour detection method provided by embodiments of the present application, Figures 9 to 11 are application scenario diagrams of a graphic contour detection method provided by embodiments of the present application. The method comprises:

[0053] In step 110, interest face data and check-in data are acquired. Each interest face in the interest face data contains a corresponding interest face contour and a geographic location identifier, and the check-in data contains a check-in geographic identifier.

[0054] For example, the interest face data and the check-in data can be front input data sources.

[0055] For example, the interest face data is Area of Interest (AOI) data, and the AOI data can be a data content in electronic map data produced by an electronic map vendor.

[0056] An Area of Interest (AOI) is a region-like geographic entity in map data, such as a cell, a school, a park, a hospital, and the like. The Area of Interest is also referred to as an entity face or an information face, and refers to a region or a region-like area of a geographic entity in a map.

[0057] For example, each interest face in the interest face data contains a corresponding interest face contour and a geographic location identifier. The Area of Interest (AOI) data is composed of basic data of the interest face contour, and each basic unit of the Area of Interest (AOI) is a binary tuple in the format of (id, boundary).

[0058] The id is a geographic location identifier of the Area of Interest (AOI), and is a unique identifier of the Area of Interest (AOI). For example, the id is generally composed of a random string, and id schemes used by various electronic map vendors can be different.

[0059] The boundary records contour coordinate string information of the interest face contour. For example, the contour coordinate string information refers to a point string of coordinate points corresponding to the interest face contour, such as [(x1, y1), (x2, y2), (x3, y3), …, (x n , y n )]. A curve drawn by the point string of coordinate points in the contour coordinate string information constitutes the interest face contour.

[0060] For example, the user shares pictures, texts and the like containing location information using various terminals (e.g., a mobile phone), which is called check-in, and the data generated thereby is called check-in data. For example, the user uses check-in or a similar function to select location sharing when publishing a video or an image in a network, and check-in data is generated.

[0061] For example, the check-in data can be location-based service (LBS) check-in data, which is data generated by the user marking a specified geographic coordinate based on LBS. The LBS check-in data can include information such as a check-in geographic coordinate at the time of check-in, a check-in geographic identifier, and a check-in object (usually a certain place POI, such as a scenic spot, a restaurant, a school, etc.).

[0062] For example, the LBS check-in data can be obtained through user positioning information reported by the user in a mobile Internet App. The LBS check-in data is a number of triplets of formats such as (id, x, y), where id is a check-in geographic identifier of a place checked in at the time of check-in, the check-in geographic identifier id is a unique identifier of the place checked in at the time of check-in, and the check-in geographic identifier id can correspond to a geographic location identifier id of an area of interest (AOI) in AOI data; x is a geographic longitude of a positioning request, and y is a geographic latitude of the positioning request. In this embodiment of the present application, the check-in data used does not require any individual identifier and does not involve privacy information. For example, the individual identifier can identify information of a mobile device or a user individual. Generally, the most original check-in data generated by a user terminal is usually collected in units of users and can have identifier information of the user. Therefore, the user terminal can remove the individual identifier from the check-in data before reporting the check-in data. In this embodiment of the present application, the check-in data obtained by the processor is check-in data that has been subjected to individual identifier removal processing in the user terminal.

[0063] It can be understood that, in the specific embodiments of the present application, if it is necessary to collect check-in data and other related data involving user information, the user's permission or consent needs to be obtained when the above embodiments of the present application are applied to specific products or technologies, and the collection, use and processing of the related data need to comply with relevant laws, regulations and standards of the country and region.

[0064] In step 120, the kernel density estimation value of each area of interest corresponding to the target location point is calculated according to the correspondence between the geographic location identifier and the check-in geographic identifier and the correspondence between the area of interest data and the check-in data.

[0065] For example, the interest surface data and the check-in data can be taken as input data, the interest surface data and the check-in data are processed by using the kernel density estimation module, and the kernel density estimation value corresponding to each interest surface in the interest surface data is calculated for the target position point.

[0066] Optionally, the check-in data further includes check-in geographic coordinates, such as Figure 3 As shown in FIG. 12, step 120 can be implemented by steps 121 to 123, specifically as follows.

[0067] In step 121, the interest surface data and the check-in data are associated according to the correspondence between the geographic location identifier and the check-in geographic identifier, to obtain a set of check-in geographic coordinates corresponding to each interest surface in the interest surface data.

[0068] In the correspondence between the geographic location identifier and the check-in geographic identifier, the geographic location identifier and the corresponding check-in geographic identifier can have the same unique identifier; or one geographic location identifier can correspond to several check-in geographic identifiers located in a preset geographic area.

[0069] For example, the check-in geographic identifiers corresponding to the geographic location identifier of the scenic spot A can include the check-in geographic identifiers of the scenic spot A, the amusement park a1 in the scenic spot A, the hotel a2 in the scenic spot A, the ball court a3 in the scenic spot A, the swimming pool a4 in the scenic spot A, and the like.

[0070] For example, when performing the kernel density estimation, the interest surface contour data and the check-in data can be associated based on the unique identifier id of the check-in location and / or the interest surface, that is, the interest surface data and the check-in data are associated based on the correspondence between the geographic location identifier and the check-in geographic identifier, and only the check-in data corresponding to the interest surface data is retained. For the set of check-in data, the unique identifier id of the check-in location and / or the interest surface can be aggregated to obtain a set of check-in geographic coordinates corresponding to the geographic location identifier of each interest surface, that is, to obtain a set of check-in geographic coordinates corresponding to each interest surface in the interest surface data. For example, the check-in data can be LBS check-in data.

[0071] In step 122, the target position point is determined according to the interest surface contour corresponding to each interest surface in the interest surface data.

[0072] Optionally, as shown in FIG. 12, step 122 can be implemented by steps 1221 to 1224, specifically as follows. Figure 4

[0073] In step 1221, the four-to-point of the interest surface contour corresponding to each interest surface in the interest surface data in the plane coordinate system is obtained.

[0074] In step 1222, a rectangular polygon is generated according to the four-to-point, to obtain the contour four-to-range of the interest surface contour corresponding to each interest surface. ​

[0075] Step 1223, according to the preset grid, the contour four-compass range is dotted, so as to obtain a plurality of position points corresponding to each contour four-compass range;

[0076] Step 1224, traversing the plurality of position points corresponding to each contour four-compass range, the target position point is determined from the plurality of position points.

[0077] For example, the four-compass point can be the four-compass point coordinate in the map, that is, the left upper longitude, the left upper latitude, the right lower longitude, and the right lower latitude. It can be understood that the four-compass point is the longitude and latitude of the four endpoints of the minimum circumscribed rectangle of the axis of a contour.

[0078] The minimum bounding rectangle (MBR) is also translated as minimum boundary rectangle, minimum containing rectangle, or minimum circumscribed rectangle. The minimum bounding rectangle refers to the maximum range of a plurality of two-dimensional shapes (such as points, straight lines, polygons) represented by two-dimensional coordinates, that is, a rectangle with a maximum horizontal coordinate, a minimum horizontal coordinate, a maximum vertical coordinate, and a minimum vertical coordinate of each vertex of the given two-dimensional shape as the boundary. Such a rectangle contains the given two-dimensional shape, and the edges are parallel to the coordinate axes.

[0079] For example, the four-compass point of the contour of each interest surface corresponding to the interest surface data in the plane coordinate system is obtained, and then a rectangular polygon is generated according to the four-compass point, so as to obtain the contour four-compass range of the contour of each interest surface corresponding to the interest surface, wherein the contour four-compass range is a rectangular polygon formed by the four-compass point.

[0080] For example, the contour four-compass range of the interest surface contour can be dotted according to the preset grid, so as to obtain a plurality of position points corresponding to each contour four-compass range. For example, in actual application, in order to reduce the calculation burden of the kernel density estimation calculation module, it is possible that each point in the plane ground surface space is not traversed, so that the preset grid (for example, 10m) with a fixed size is dotted, and the range of the dot traversal is usually rectangular. For example, when dotting, the contour four-compass range of the interest surface contour can be traversed according to the preset grid, and the minimum longitude Min_Lon, the maximum longitude Max_Lon, the minimum latitude Min_Lat, and the maximum latitude Max_lat in the four-compass point of the four-compass point contour corresponding to the geographic location identifier id of the interest surface are recorded. The four-compass point of the four-compass point contour includes the left upper longitude, the left upper latitude, the right lower longitude, and the right lower latitude. Specifically, the minimum longitude Min_Lon, the maximum longitude Max_Lon, the minimum latitude Min_Lat, and the maximum latitude Max_lat in the four-compass point are determined according to the specific geographic location of the four-compass point contour.

[0081] For example, the outline can be expanded outwards relative to its boundaries based on custom parameters to form an expanded outline boundary. Points can be marked on the expanded outline boundary based on a preset grid to obtain multiple location points corresponding to each expanded outline boundary. For example, the range of point traversal can be a longitude interval [Min_Lon–b, Max_Lon+b] and a latitude interval [Min_Lat–b, Max_Lat+b], where b is a custom parameter representing the distance the traversal range extends outwards relative to the outline boundary. For example, b can be any value between 5% and 10% of the expanded outline boundary, such as b being 5% or 10% of the expanded outline boundary.

[0082] For example, multiple location points corresponding to the boundaries of each contour can be traversed, and the target location point for kernel density estimation can be determined from the multiple location points. Then, for each target location point, the kernel density estimate value of the target location point for the interest surface can be calculated in the manner described above.

[0083] Step 123: Based on the check-in geographic coordinate set, calculate the kernel density estimate of the target location point corresponding to each interest surface in the interest surface data.

[0084] Optional, such as Figure 5 As shown, step 123 can be implemented through steps 1231 to 1234, specifically as follows:

[0085] Step 1231: Based on the check-in geographic coordinate set, determine the target number of check-in geographic coordinates corresponding to each interest surface in the interest surface data.

[0086] Step 1232: Obtain the target distance from the check-in geographic coordinates to the target location point for each interest surface in the interest surface data.

[0087] Step 1233: Obtain the target bandwidth corresponding to the target kernel function.

[0088] Step 1234: Based on the number of targets, target distance, target bandwidth and target kernel function, calculate the kernel density estimate of each interest surface in the interest surface data corresponding to the target location point. The kernel density estimate is used to characterize the intensity of the check-in behavior of the corresponding target location point.

[0089] For example, for each interest surface and its corresponding set of check-in geographic coordinates, the kernel density can be estimated spatially using the following formula (1). The kernel density estimate represents the intensity of the check-in behavior and reflects the extent to which the check-in geographic coordinates in the set belong to the corresponding interest surface. Specifically, for each target location point x in space, the kernel density estimate corresponding to the interest surface at that target location point is:

[0090]

[0091] wherein, n is the target number of check-in geographic coordinates corresponding to the interest surface; r is the target distance from the check-in geographic coordinate to the target position point x; K is the target kernel function corresponding to the target bandwidth h, wherein the target bandwidth h>0.

[0092] wherein, the target number of check-in geographic coordinates corresponding to each interest surface in the interest surface data can be determined according to the check-in geographic coordinate set.

[0093] wherein, the main factors affecting the kernel density estimation effect are the mathematical form of the target kernel function K and the target bandwidth h, wherein the commonly used target kernel function K can include the forms of uniform kernel function (Uniform), triangle kernel function (Triangle), Gaussian kernel function (Gaussian), quartic kernel function (Epanechnikov), cosine kernel function (Cosine) and the like.

[0094] For example, in an embodiment of the embodiment of the present application, the target kernel function can be fixed as the quartic kernel function, and the bandwidth h is set as a self-defined and modifiable hyperparameter of the model proposed in the embodiment of the present application. The formula of the quartic kernel function can be expressed as the following formula (2):

[0095]

[0096] wherein, the function K(u) corresponds to the function K in the above formula (1) h , and u represents r / h in the above formula, wherein -1≤u≤1.

[0097] wherein, the different target kernel functions K have little influence on the asymptotic property of the kernel density estimation, and therefore the selection of the target bandwidth h becomes the focus of the research on the kernel density estimation. A large bandwidth can cause over-smoothing, which can cover the real structure of the data; a small bandwidth can produce a sharp multi-peak density estimation (similar to the sharp edges of a triangle); therefore, the target of the embodiment of the present application is to find an optimal bandwidth as the target bandwidth, so that the error between the kernel density estimation value calculated by using the determined target bandwidth and the real density is minimized.

[0098] Optionally, as shown in Figure 6 , the step 1233 can be implemented by the step 12331 to the step 12336, specifically:

[0099] The step 12331 searches a plurality of candidate bandwidths in a first preset interval according to a preset interval.

[0100] Step 12332, based on each candidate bandwidth, calculate the first kernel density estimation value corresponding to each interest surface in the target position point in the interest surface data;

[0101] Step 12333, generate the first contour line of the first kernel density estimation value, and generate the first isosurface according to the first contour line;

[0102] Step 12334, determine the first target isosurface according to the spatial projection area of the first isosurface and the spatial projection area of the interest surface profile corresponding to the first isosurface;

[0103] Step 12335, calculate the first intersection-over-union index between the profile of the first target isosurface corresponding to each interest surface in the interest surface data and the interest surface profile;

[0104] Step 12336, traverse the plurality of candidate bandwidths, and determine the candidate bandwidth corresponding to the highest value in the average value of the first intersection-over-union index of all interest surfaces in the interest surface data as the target bandwidth corresponding to the target kernel function.

[0105] In the process of graph contour detection, the largest uncertain condition is the self-defined parameter bandwidth h in the kernel density estimation module. Different selection of bandwidth h will generate different kernel density estimation results, thereby generating different contour lines and isosurfaces, which directly affects the error evaluation of each interest surface. Therefore, the best bandwidth needs to be selected as the target bandwidth for kernel density estimation.

[0106] For example, it is assumed that most of the interest surface profiles and the target isosurface profiles (such as the best isosurface profile) should be relatively similar, that is, the target isosurface profile (such as the best isosurface profile) representing the virtual cognitive space obtained by mining the spatial cognitive behavior (such as LBS check-in) should be relatively close to the interest surface profile in the geographic data in most cases, that is, has a relatively high IOU index.

[0107] Therefore, the average value of the IOU index in the error analysis of all interest surfaces can be used as the optimization target of the bandwidth h parameter. For example, the optimization method can use the grid search method, that is, searching for the best bandwidth h at a predetermined interval within a certain interval. If the best bandwidth h is found, the iteration is stopped. For example, the predetermined interval of the bandwidth can be selected within 100m-10000m, and in order to reduce the calculation cost, it can be selected within 100m-2000m. For example, the bandwidth can be searched at a predetermined interval of 100m within the predetermined interval of 100m-2000m, such as 100m, 200m, 300m, …, 1800m, 1900m, 2000m. The smaller the search interval, the more accurate the best bandwidth obtained by searching, but the larger the calculation amount, so the benefits and costs can be balanced to select a suitable predetermined interval.

[0108] For example, there can be two searching schemes:

[0109] 1) All the interest surfaces share the same bandwidth h parameter, and the goal is to make the average value of the IOU indicators of all the interest surfaces highest. Specifically, a plurality of candidate bandwidths are searched in a first preset interval according to a preset interval, and all the interest surfaces in the interest surface data to be processed share the same bandwidth parameter; based on each candidate bandwidth, a first kernel density estimation value of the target position point corresponding to each interest surface in the interest surface data is calculated, a first contour line of the first kernel density estimation value is generated, and a first isosurface is generated according to the first contour line, then the first target isosurface is determined according to the spatial projection area of the first isosurface and the spatial projection area of the interest surface profile corresponding to the first isosurface, then the first intersection-over-union indicator between the profile of the first target isosurface corresponding to each interest surface in the interest surface data and the interest surface profile is calculated, and then the plurality of candidate bandwidths are traversed, and the candidate bandwidth corresponding to the highest value in the average value of the first intersection-over-union indicators of all the interest surfaces in the interest surface data is determined as the target bandwidth corresponding to the target kernel function.

[0110] 2) The optimal bandwidth h can be searched according to the interest surfaces of different categories (such as hospitals, schools, residential areas, commercial buildings, public places, etc.), and the interest surfaces of the same category share the same bandwidth h parameter, and the goal is to make the average value of the IOU indicators of the interest surfaces under the same category highest. The assumption here is that different types of interest surfaces have different geographic spatial recognition rules. Specifically, a plurality of candidate bandwidths are searched in a first preset interval according to a preset interval, and the interest surfaces of the same category in the interest surface data to be processed share the same bandwidth parameter; based on each candidate bandwidth, a first kernel density estimation value of the target position point corresponding to each interest surface in the interest surface data under the same category is calculated, a first contour line of the first kernel density estimation value is generated, and a first isosurface is generated according to the first contour line, then the first target isosurface is determined according to the spatial projection area of the first isosurface and the spatial projection area of the interest surface profile corresponding to the first isosurface, then the first intersection-over-union indicator between the profile of the first target isosurface corresponding to each interest surface in the interest surface data under the same category and the interest surface profile is calculated, and then the plurality of candidate bandwidths are traversed, and the candidate bandwidth corresponding to the highest value in the average value of the first intersection-over-union indicators of all the interest surfaces under the same category in the interest surface data is determined as the target bandwidth of the target kernel function corresponding to the category. All categories in the interest surface data are traversed to obtain the target bandwidth of the target kernel function corresponding to different categories.

[0111] For example, the geospatial cognitive rule is usually used to express the geospatial cognition. The geospatial cognition is to study the behavior process of understanding the mutual position, spatial distribution, dependent relationship and change rule of various things and phenomena on the earth surface. The geospatial cognition is the information processing process of the organic connection of perception, attention, image, memory, learning, thinking, language, concept formation, problem solving, emotion and personality difference. The medium of the geospatial cognition is usually the map or image (including digital form and analog form) describing the geographical environment and phenomenon, and therefore is closely related to the map spatial cognition. The map spatial cognition is divided into the cognition of the map editor and designer and the cognition of the map user. Among them, the cognition of the map editor emphasizes the cognition of the expressed things and phenomena and the expression form of the expression content, so as to establish the image and symbolized summary model of the objective world; the cognition of the map user is to combine the spatial knowledge of the user and indirectly cognize the real world according to the map. For example, generally, when a user A selects to check in at a distance of 200 m from a park, the user A may select the geographical position coordinate of the check-in as the park, which means that in the geospatial cognition of the user A, the position where the user A is currently located can be replaced by the park. For example, a user B may only select the geographical position coordinate of the check-in as a convenience store when the user B is at a distance (for example, several meters) close to the convenience store. The above is the difference in the geospatial cognition of the different categories of the interest surface (geographical entity), which directly affects the spatial distribution of the check-in data and also affects the best bandwidth calculated by the kernel density estimation. Therefore, the different categories of the interest surface generally have different best bandwidths.

[0112] In step 130, contour lines of the kernel density estimation value are generated, and a plurality of contour surfaces are generated according to the contour lines.

[0113] Optionally, as shown in FIG. 13, step 130 can be implemented by steps 131 to 132, specifically: Figure 7

[0114] In step 131, according to the contour line generation algorithm, points with the same value of the kernel density estimation value are connected into a curve to generate a plurality of contour lines, wherein the spatial contour formed by the combination of each contour line and the rectangular boundary of the contour four-to-range is a closed ring line.

[0115] For example, when the contour line corresponding to the kernel density estimation value is drawn, for a certain interest surface, based on the kernel density estimation value of each position point of the space generated by the last step of traversal, the contour line contour of the interest surface is generated according to the contour line generation algorithm.

[0116] ​The contour line is a line connecting contour points in a raster dataset representing a continuous phenomenon, such as elevation, temperature, rainfall, pollution level, or atmospheric pressure. These line features connect together the pixels in the input that have the same constant value. The distribution of the contour lines can show the variation of the value across the surface. The smaller the variation of the value, the larger the spacing of the lines. The faster the value rises or falls, the smaller the spacing of the lines.

[0117] For example, the contour line generation algorithm can be implemented by the following steps: 1) finding all squares formed by four adjacent data points in the data grid; 2) determining the relationship between the four data points and the threshold value to generate 01 code; 3) determining the topological relationship between the contour line and the four data points according to the relationship of the code generated in the previous step; 4) determining the intersection of the contour line and the square by linear interpolation method according to the topological relationship; and 5) sequentially connecting the contour line segments to obtain the contour line.

[0118] In step 132, the closed polygon formed inside the ring line corresponding to each contour line in the plurality of contour lines is taken as the contour surface to obtain a plurality of contour surfaces.

[0119] The contour line is a curve formed by connecting points with equal kernel density estimation values in space. Different intervals of the kernel density estimation value can be selected to generate contour lines, and thus there can be multiple contour lines. The contour line and the rectangular boundary formed by the traversal range form a closed ring (LinearRing), and the ring forms a closed polygon inside the ring, which can be referred to as a contour surface.

[0120] For example, the height of the ground is continuous, but different intervals can be taken, such as 100 m contour lines and 110 m contour lines. Thus, a plurality of contour lines can be obtained by taking intervals of 10 m. Similarly, the value of the contour line can have different numerical meanings, such as altitude in the contour line. In the embodiments of the present application, the value in the contour line is the kernel density estimation value, which can be taken at intervals of 0.1 or 0.01. The interval can be determined according to the range of the kernel density estimation value. Alternatively, the kernel density estimation value can be taken at intervals of N segments after taking the logarithm of the minimum value to the maximum value. N can be 20, 50, 100, etc.

[0121] In step 140, the target contour surface corresponding to each interest surface in the interest surface data is determined from the plurality of contour surfaces.

[0122] Alternatively, the target contour surface corresponding to each interest surface in the interest surface data is determined from the plurality of contour surfaces, including: determining the target contour surface corresponding to each interest surface in the interest surface data according to the spatial projection area of the contour surface and the spatial projection area of the contour surface corresponding to the interest surface contour.

[0123] Optionally, according to the spatial projection area of the isosurface and the spatial projection area of the contour of the interest surface corresponding to the isosurface, a target isosurface corresponding to each interest surface in the interest surface data is determined, comprising:

[0124] The spatial projection area of each isosurface and the spatial projection area of the contour of the interest surface corresponding to each isosurface are calculated.

[0125] The isosurface with the spatial projection area closest to the spatial projection area of the contour of the target interest surface is selected from the spatial projection area of the isosurfaces as the target isosurface corresponding to the target interest surface.

[0126] For example, for the contours of the plurality of isosurfaces generated in the previous step, the spatial projection area of each isosurface is calculated in turn, and the isosurface with the spatial projection area closest to the spatial projection area of the contour of the corresponding interest surface is selected from the plurality of isosurfaces as the best isosurface. The determined best isosurface is taken as the target isosurface. It can be considered that, through the LBS check-in data, the best isosurface reflects the cognitive contour of the user group of the mobile Internet in the geographical space cognition to the ground object corresponding to the interest surface.

[0127] For example, the spatial projection area can be an area obtained by projection conversion based on a geographic coordinate system (latitude and longitude). For example, the contour of the isosurface is subjected to Gauss projection or Albers projection and the like based on the geographic coordinate system to obtain the spatial projection area of the isosurface. For example, the contour of the interest surface is subjected to Gauss-Kruger projection or Albers projection and the like based on the geographic coordinate system to obtain the spatial projection area of the contour of the interest surface.

[0128] For example, Gauss projection, also known as Gauss-Kruger projection, is a kind of orthomorphic projection between the earth ellipsoid and the plane. It is assumed that an elliptical cylindrical surface is horizontally wrapped outside the earth ellipsoid, and is tangent to a certain meridian (this meridian is called central meridian or axis meridian), and the center axis of the elliptical cylinder passes through the center of the ellipsoid. Then, a certain projection method is used to project the regions on both sides of the central meridian within a certain range of meridional difference to the elliptical cylindrical surface, and then the cylindrical surface is unfolded to become the projection surface, which is the Gauss projection.

[0129] For example, Albers projection, also known as "orthaxial equal-area conic projection" and "double-standard latitude equal-area conic projection", is a kind of conic projection. The latitudes are concentric circular arcs, and the meridians are radii of a circle, and the meridian angle is proportional to the corresponding meridional difference. The two cutting latitudes are not deformed after projection. The area of the projection region remains equal to the actual area.

[0130] Step 150, detecting an abnormal interest surface profile according to the intersection-union ratio index between the profile of the target isosurface corresponding to each interest surface in the interest surface data and the interest surface profile.

[0131] Optionally, as shown in Figure 8 Step 150 can be implemented through steps 151 to 152, specifically:

[0132] Step 151, calculating the intersection-union ratio index between the profile of the target isosurface corresponding to each interest surface in the interest surface data and the interest surface profile.

[0133] Optionally, calculating the intersection-union ratio index between the profile of the target isosurface corresponding to each interest surface in the interest surface data and the interest surface profile includes:

[0134] Obtaining an intersection operator between the profile of the target isosurface corresponding to each interest surface in the interest surface data and the interest surface profile;

[0135] Obtaining a union operator between the profile of the target isosurface corresponding to each interest surface in the interest surface data and the interest surface profile;

[0136] Determining a first area operator corresponding to the intersection operator and a second area operator corresponding to the union operator;

[0137] According to the ratio of the first area operator and the second area operator, determining the intersection-union ratio index between the profile of the target isosurface corresponding to each interest surface in the interest surface data and the interest surface profile.

[0138] GIS spatial analysis refers to the implementation of analyzing spatial data in GIS (Geographic Information System), that is, obtaining and analyzing information about the spatial position, distribution, shape, formation and evolution of geographic objects from spatial data.

[0139] Intersection is a GIS spatial analysis operator that calculates the intersection part of two profiles, as shown in Figure 9 The input is profile A and profile B, the intersection part of the two profiles is C, and the output is C.

[0140] Union is a GIS spatial analysis operator that calculates the union part of two profiles, as shown in Figure 10 The input is profile A and profile B, the union part of the two profiles is A+B, and the output is A+B.

[0141] For example, for the profile of the interest surface and the profile of the target contour surface (best contour surface) of a certain interest surface, the intersection-over-union (IOU) evaluation index can be calculated, and the calculation formula can be expressed as formula (3) as follows:

[0142] IOU = Area (Intersection (AOI, CS)) / Area (Union (AOI, CS)) * 100% (3);

[0143] Wherein, AOI represents the profile of the interest surface (AOI) corresponding to the interest surface id, and CS represents the profile of the selected target contour surface (Contour Surface); Intersection (AOI, CS) is an intersection operator in GIS spatial analysis, which represents the intersection part of the profile of the interest surface (AOI) and the profile of the target contour surface (CS), and Union (AOI, CS) is a union operator in GIS spatial analysis, which represents the union part of the profile of the interest surface (AOI) and the profile of the Contour Surface (CS). Area (x) is an area operator in GIS spatial analysis, which represents the area of the profile x, and the areas of the profiles should be calculated under the definition of the same projection coordinate system. In the embodiment of the present application, Area (Intersection (AOI, CS)) represents the area of the intersection part of the profile of the interest surface (AOI) and the profile of the Contour Surface (CS), and Area (Union (AOI, CS)) represents the area of the union part of the profile of the interest surface (AOI) and the profile of the Contour Surface (CS).

[0144] Wherein, the value of the IOU index is 0% to 100%, generally, the higher the IOU index, the smaller the difference between the two profiles, IOU = 100% represents that the two profiles are completely consistent, on the contrary, the lower the IOU index, the greater the difference between the two profiles, IOU = 0% represents that the two profiles have no overlap. Therefore, the IOU index is used as the result of error analysis.

[0145] Step 152, arrange the intersection-over-union index from small to large, and sequentially extract the interest surface profile corresponding to the top k% data in the arranged intersection-over-union index from the interest surface data, and determine the abnormal interest surface profile, wherein k% is the extraction ratio.

[0146] The one or more target bandwidths (optimal bandwidth h parameters) searched in the last step are sequentially subjected to kernel density estimation, contour generation, target contour surface (optimal contour surface) selection and error analysis, and finally the IOU index for error evaluation and analysis is calculated on the basis of each interest surface in its corresponding target bandwidth (optimal bandwidth parameter). The IOU index corresponding to each interest surface is arranged from small to large, and the top k% of the data corresponding to the interest surface profile is determined as the abnormal interest surface profile, wherein k% is the extraction ratio. For example, the k% can be 5%. The specific value of k% can be determined according to the business requirements. For example, if a wide and large-scale screening is required, the interest surface profile corresponding to the top 5% of the data can be considered as the abnormal interest surface profile. For example, if a small-cost and small-scale investigation is required, the interest surface profile corresponding to the top 0.1% of the data can be considered as the abnormal interest surface profile.

[0147] The embodiments of the present application finally obtain the error evaluation result according to the IOU index of the optimal contour surface and the interest surface profile through the steps of kernel density estimation, contour generation, contour surface mining, optimal contour surface selection and error analysis, and optimal bandwidth selection. The IOU index representing the error situation is sorted from small to large, and the automatic preliminary screening detection list of the abnormal interest surface profile is obtained.

[0148] For example, Figure 11 is a graphical diagram. The L-shaped profile 1101 is the actual interest surface profile, and the gray scale background 1102 in the bottom diagram is the result of kernel density estimation corresponding to the interest surface. The light gray color (such as 1103 in the figure) represents a low kernel density estimation value, and the dark gray color (such as 1104 in the figure) represents a high kernel density estimation value. The ring line 1105 is the contour of the kernel density estimation, and the closed contour inside can form a contour surface. Figure 11 No matter which contour surface is selected, the IOU index with the interest surface profile is always low, which can reflect that at least from the result of spatial cognition, there is a potential problem of too small profile of the interest surface.

[0149] Optionally, as shown in Figure 12 , the step 150 can further include steps 160 to 180, specifically:

[0150] Step 160, sending the abnormal interest surface profile to the target terminal;

[0151] Step 170, receiving the cause analysis result of the abnormal interest surface profile fed back by the target terminal;

[0152] Step 180, updating the interest surface data according to the cause analysis result.

[0153] For example, the processor can send the abnormal interest surface profile to a target terminal, where sending the abnormal interest surface profile can be sending an automated preliminary screening detection list containing the abnormal interest surface profile, so that the target terminal displays the abnormal interest surface profile, and the target terminal receives a user input of a cause analysis result for the abnormal interest surface profile; then, the processor receives the cause analysis result for the abnormal interest surface profile fed back by the target terminal, and sends the cause analysis result to a client generating the interest surface data, so that the client updates the interest surface data according to the cause analysis result.

[0154] For example, the cause of the abnormal interest surface can include:

[0155] a) Interest surface profile deviation; assuming a scenario as follows: the interest surface profile of Laoshan Scenic Area in Qingdao, Shandong is mistakenly marked to Yantai, but the mobile Internet user group still performs LBS check-in behavior in Laoshan Scenic Area in Qingdao, at this time, the interest surface profile does not coincide with the best isosurface profile, and the IOU index is 0%, which is a typical abnormal interest surface.

[0156] b) Incorrect spatial range of interest surface profile; assuming a scenario as follows: the interest surface profile of a certain hospital is drawn too large, so that the residential area near the hospital is also included in the interest surface profile of the hospital. However, in the LBS check-in behavior, users will only check in near the hospital, rarely check in within the residential area, so the check-in geographic coordinates are mainly concentrated in the hospital, the interest surface profile has low coincidence with the best isosurface profile, and the calculated IOU index is low.

[0157] c) Incorrect spatial cognition; that is, the interest surface profile itself is correct, but the spatial cognition of the crowd to the place has a large deviation, resulting in low coincidence of the interest surface profile with the best isosurface profile.

[0158] Wherein, the embodiments of the present application can effectively screen out potential abnormal objects from a large number of interest surfaces, realize the identification of the abnormal interest surface profile, greatly save the operation cost and speed up the mining and correction speed of the abnormal interest surface profile. Wherein, after detecting the abnormal interest surface profile, the abnormal interest surface profile needs to be sent to a target terminal for manual discrimination, that is, the above three abnormal situations still need to be manually discriminated on the basis of the automated preliminary screening detection list of the abnormal interest surface profile, and users can analyze the cause of the abnormal interest surface profile according to the fact objective situation.

[0159] All the above technical solutions can be combined in any way to form optional embodiments of the present application, which will not be repeated here.

[0160] The embodiment of the present application obtains interest surface data and check-in data, each interest surface in the interest surface data contains a corresponding interest surface contour and a geographic location identifier, and the check-in data contains a check-in geographic identifier; according to the corresponding relationship between the geographic location identifier and the check-in geographic identifier and the interest surface data and the check-in data, the kernel density estimation value of a target position point corresponding to each interest surface in the interest surface data is calculated; an isogram of the kernel density estimation value is generated, and a plurality of isosurfaces are generated according to the isogram; a target isosurface corresponding to each interest surface in the interest surface data is determined from the plurality of isosurfaces; and the intersection and union ratio index between the contour of the target isosurface corresponding to each interest surface in the interest surface data and the interest surface contour is detected. The embodiment of the present application can realize preliminary automatic screening, reduce the labor operation cost of abnormal interest surface contour detection, speed up the data optimization efficiency, and improve the accuracy of abnormal interest surface contour detection by calculating the kernel density estimation value of each interest surface based on the interest surface data and the check-in data, calculating the isogram and the isosurface of the kernel density estimation value, and detecting the intersection and union ratio index between the contour of the target isosurface and the interest surface contour.

[0161] In addition, the embodiment of the present application can also automatically color the original image based on the color matching model trained based on the sample data with color matching annotation information, without manually determining the color matching effect, thereby improving the efficiency of the graphic contour detection and realizing batch automatic production.

[0162] In order to better implement the graphic contour detection method of the embodiment of the present application, the embodiment of the present application further provides a graphic contour detection device. Please refer to Figure 13 , Figure 13 The first structural schematic diagram of the graphic contour detection device provided by the embodiment of the present application is shown in FIG. 2. The graphic contour detection device 200 can include:

[0163] The obtaining module 210 is configured to obtain interest surface data and check-in data, each interest surface in the interest surface data contains a corresponding interest surface contour and a geographic location identifier, and the check-in data contains a check-in geographic identifier;

[0164] The calculation module 220 is configured to calculate the kernel density estimation value of a target position point corresponding to each interest surface in the interest surface data according to the corresponding relationship between the geographic location identifier and the check-in geographic identifier and the interest surface data and the check-in data;

[0165] The generation module 230 is configured to generate an isogram of the kernel density estimation value, and generate a plurality of isosurfaces according to the isogram;

[0166] The determination module 240 is configured to determine a target isosurface corresponding to each interest surface in the interest surface data from the plurality of isosurfaces;

[0167] The detection module 250 is configured to detect an abnormal interest surface profile according to an intersection-union ratio index between a profile of a target equipotential surface corresponding to each interest surface in the interest surface data and the interest surface profile.

[0168] Optionally, the calculation module 220 comprises:

[0169] The association unit is configured to associate the interest surface data and the check-in data according to a corresponding relationship between the geographic position identifier and the check-in geographic identifier, to obtain a check-in geographic coordinate set corresponding to each interest surface in the interest surface data.

[0170] The first determination unit is configured to determine a target position point according to an interest surface profile corresponding to each interest surface in the interest surface data.

[0171] The first calculation unit is configured to calculate a kernel density estimation value of the target position point corresponding to each interest surface in the interest surface data according to the check-in geographic coordinate set.

[0172] Optionally, the first calculation unit can be configured to: determine a target number of check-in geographic coordinates corresponding to each interest surface in the interest surface data according to the check-in geographic coordinate set; obtain a target distance from the check-in geographic coordinates corresponding to each interest surface in the interest surface data to the target position point; obtain a target bandwidth corresponding to a target kernel function; and calculate the kernel density estimation value of the target position point corresponding to each interest surface in the interest surface data according to the target number, the target distance, the target bandwidth and the target kernel function, the kernel density estimation value being used to represent the intensity of the check-in behavior of the corresponding target position point.

[0173] Optionally, when obtaining the target bandwidth corresponding to the target kernel function, the first calculation unit can be configured to: search a plurality of candidate bandwidths in a first preset interval according to a preset interval; calculate a first kernel density estimation value of the target position point corresponding to each interest surface in the interest surface data based on each candidate bandwidth; generate a first contour line of the first kernel density estimation value, and generate a first equipotential surface according to the first contour line; determine a first target equipotential surface according to a spatial projection area of the first equipotential surface and a spatial projection area of the interest surface profile corresponding to the first equipotential surface; calculate a first intersection-union ratio index between a profile of the first target equipotential surface corresponding to each interest surface in the interest surface data and the interest surface profile; and traverse the plurality of candidate bandwidths, and determine a candidate bandwidth corresponding to a highest value in an average value of the first intersection-union ratio indexes of all the interest surfaces in the interest surface data as the target bandwidth corresponding to the target kernel function.

[0174] Optionally, the first determining unit can be configured to: acquire four-corner points of each interest surface profile in the interest surface data in a plane coordinate system; generate a rectangular polygon according to the four-corner points to obtain a profile four-corner range of each interest surface profile; dot the profile four-corner range according to a preset grid to obtain a plurality of position points corresponding to each profile four-corner range; and traverse the plurality of position points corresponding to each profile four-corner range to determine a target position point from the plurality of position points.

[0175] Optionally, the generating module 230 comprises:

[0176] The first generating unit is configured to connect points with the same value in the kernel density estimation value into a curve according to an isogram generation algorithm to generate a plurality of isograms, wherein a space profile formed by combination of each isogram and a rectangular boundary of the profile four-corner range is a closed loop line.

[0177] The second generating unit is configured to form a closed polygon inside the loop line corresponding to each isogram in the plurality of isograms as an isosurface to obtain a plurality of isosurfaces.

[0178] Optionally, the determining module 240 can be configured to determine a target isosurface corresponding to each interest surface in the interest surface data according to a space projection area of the isosurface and a space projection area of the interest surface profile corresponding to the isosurface.

[0179] Optionally, when determining the target isosurface corresponding to each interest surface in the interest surface data according to the space projection area of the isosurface and the space projection area of the interest surface profile corresponding to the isosurface, the determining module 240 can be configured to: calculate the space projection areas of the plurality of isosurfaces and the space projection areas of the interest surface profiles corresponding to each isosurface in the plurality of isosurfaces; and select an isosurface with a space projection area closest to the space projection area of the interest surface profile of the target interest surface from the space projection areas of the plurality of isosurfaces as the target isosurface corresponding to the target interest surface.

[0180] Optionally, the detecting module 250 comprises:

[0181] The second calculating unit is configured to calculate an intersection-union ratio index between a profile of the target isosurface corresponding to each interest surface in the interest surface data and the interest surface profile.

[0182] The second determining unit is configured to arrange the intersection-union ratio indexes from small to large, sequentially extract interest surface profiles corresponding to data ranked in the top k% from the arranged intersection-union ratio indexes from the interest surface data, and determine the interest surface profiles as abnormal interest surface profiles, wherein k% is an extraction ratio.

[0183] Optionally, in the calculation of the intersection-union ratio index between the contour of the target isosurface corresponding to each interest surface in the interest surface data and the contour of the interest surface, the second calculation unit is specifically configured to: obtain an intersection operator between the contour of the target isosurface corresponding to each interest surface in the interest surface data and the contour of the interest surface; obtain a joint operator between the contour of the target isosurface corresponding to each interest surface in the interest surface data and the contour of the interest surface; determine a first area operator corresponding to the intersection operator and a second area operator corresponding to the joint operator; and determine the intersection-union ratio index between the contour of the target isosurface corresponding to each interest surface in the interest surface data and the contour of the interest surface according to a ratio of the first area operator to the second area operator.

[0184] Optionally, the graphic contour detection apparatus 200 further includes:

[0185] The sending module is configured to send the abnormal interest surface contour to the target terminal.

[0186] The receiving module is configured to receive a cause analysis result for the abnormal interest surface contour fed back by the target terminal.

[0187] The updating module is configured to update the interest surface data according to the cause analysis result.

[0188] The above-described various units in the graphic contour detection apparatus can be all or partially implemented by software, hardware, and a combination thereof. The above-described various units can be embedded in or independent of a processor in a computer device in a hardware form, or can be stored in a memory in the computer device in a software form, so as to be called and executed by a processor to perform operations corresponding to the above-described various units.

[0189] The graphic contour detection apparatus 200 can be integrated in a terminal or a server with a storage and a processor installed and having a computing capability, or the graphic contour detection apparatus 200 is the terminal or the server.

[0190] Optionally, the present application further provides a computer device including a memory and a processor, and the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described various method embodiments.

[0191] Figure 14 A structural schematic diagram of a computer device provided by the embodiments of the present application is shown in FIG. 5. The computer device can be a terminal or a server as shown in FIG. 5. As shown in FIG. 5, the computer device includes a processor 501 and a memory 502. Figure 1 The processor 501 is configured to execute a computer program stored in the memory 502 to implement the steps in the above-described various method embodiments. Figure 14As shown, the computer device 300 can include a communication interface 301, a memory 302, a processor 303 and a communication bus 304. The communication interface 301, the memory 302 and the processor 303 can communicate with each other through the communication bus 304. The communication interface 301 is configured to perform data communication between the apparatus 700 and an external device. The memory 302 can be configured to store software programs and modules. The processor 303 can execute the software programs and modules stored in the memory 302, for example, the software programs corresponding to the operations in the foregoing method embodiments.

[0192] Optionally, the processor 303 can invoke the software programs and modules stored in the memory 302 to perform the following operations: obtaining interest face data and check-in data, each interest face in the interest face data containing a corresponding interest face contour and a geographic location identifier, and the check-in data containing a check-in geographic identifier; calculating, according to a correspondence between the geographic location identifier and the check-in geographic identifier and the interest face data and the check-in data, a kernel density estimation value of a target location point corresponding to each interest face in the interest face data; generating an isogram of the kernel density estimation value, and generating a plurality of isosurfaces according to the isogram; determining, from the plurality of isosurfaces, a target isosurface corresponding to each interest face in the interest face data; and detecting an abnormal interest face contour according to an intersection-union ratio index between a contour of the target isosurface corresponding to each interest face in the interest face data and the interest face contour.

[0193] The present application also provides a computer readable storage medium for storing a computer program. The computer readable storage medium can be applied to a computer device, and the computer program causes the computer device to perform the corresponding procedures in the graph contour detection method in the embodiments of the present application. For brevity, details are not repeated here.

[0194] The present application also provides a computer program product, which includes computer instructions stored in a computer readable storage medium. The processor of a computer device reads the computer instructions from the computer readable storage medium, and the processor executes the computer instructions, so that the computer device performs the corresponding procedures in the graph contour detection method in the embodiments of the present application. For brevity, details are not repeated here.

[0195] The present application also provides a computer program, which includes computer instructions stored in a computer readable storage medium. The processor of a computer device reads the computer instructions from the computer readable storage medium, and the processor executes the computer instructions, so that the computer device performs the corresponding procedures in the graph contour detection method in the embodiments of the present application. For brevity, details are not repeated here.

[0196] It should be understood that the processor of the embodiments of the present application can be an integrated circuit chip with a processing capability of signals. In the implementation process, each step of the method embodiments described above can be completed by the integrated logic circuit of hardware in the processor or the instructions in the form of software. The processor described above can be a general processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, a discrete gate or transistor logic device, a discrete hardware component. The disclosed methods, steps and logic block diagrams in the embodiments of the present application can be implemented or executed. The general processor can be a microprocessor or the processor can also be any conventional processor or the like. The steps of the method disclosed in conjunction with the embodiments of the present application can be directly embodied as a hardware code processor for execution, or can be executed by a combination of hardware and software modules in the code processor. The software module can be located in a random memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register or other mature storage medium in the art. The storage medium is located in the memory, and the processor reads the information in the memory, and combines the hardware to complete the steps of the above method.

[0197] It is to be understood that the memory in the embodiments of the present application can be a volatile memory or a nonvolatile memory, or can include both volatile and nonvolatile memory. Among them, the nonvolatile memory can be a read-only memory (Read-Only Memory, ROM), a programmable read-only memory (Programmable ROM, PROM), an erasable programmable read-only memory (Erasable PROM, EPROM), an electrically erasable programmable read-only memory (Electrically EPROM, EEPROM) or a flash memory. The volatile memory can be a random access memory (Random Access Memory, RAM) used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as static random access memory (Static RAM, SRAM), dynamic random access memory (Dynamic RAM, DRAM), synchronous dynamic random access memory (Synchronous DRAM, SDRAM), double data rate synchronous dynamic random access memory (Double Data Rate SDRAM, DDR SDRAM), enhanced synchronous dynamic random access memory (Enhanced SDRAM, ESDRAM), synchronous link dynamic random access memory (Synchlink DRAM, SLDRAM) and direct memory bus random access memory (Direct Rambus RAM, DR RAM). It should be noted that the memory of the system and method described herein is intended to include, but not limited to, these and any other suitable types of memory.

[0198] It should be understood that the above-mentioned memory is exemplary but not limiting, for example, the memory in the embodiments of the present application can also be static random access memory (static RAM, SRAM), dynamic random access memory (dynamic RAM, DRAM), synchronous dynamic random access memory (synchronous DRAM, SDRAM), double data rate synchronous dynamic random access memory (double data rate SDRAM, DDR SDRAM), enhanced synchronous dynamic random access memory (enhanced SDRAM, ESDRAM), synchronous link dynamic random access memory (synch link DRAM, SLDRAM) and direct memory bus random access memory (Direct Rambus RAM, DR RAM) and the like. That is, the memory in the embodiments of the present application is intended to include, but not limited to, these and any other suitable types of memory.

[0199] Those skilled in the art can clearly understand that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0200] Those skilled in the art can clearly understand that, for the convenience and brevity of the description, the specific working processes of the above-described system, device and unit can refer to the corresponding processes in the foregoing method embodiments, which will not be repeated here.

[0201] In several embodiments provided in the present application, it should be understood that the disclosed system, device and method can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division, and there can be another division manner in actual implementation. For example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interface, device or unit, and can be electrical, mechanical or other forms.

[0202] The units described as separate components can or can not be physically separate, and the components shown as units can or can not be physical units, i.e. they can be located in one place or distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the embodiment.

[0203] In addition, each functional unit in the embodiments of the present application can be integrated into a processing unit, or each unit can exist physically, or two or more units can be integrated into one unit.

[0204] If the functions are realized in the form of software function units and sold or used as independent products, they can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application essentially or the parts that make contributions to the prior art or parts of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a number of instructions for making a computer device (which can be a personal computer, a server) execute all or part of the steps of the methods described in the embodiments of the present application. The foregoing storage medium includes: U disk, mobile hard disk, ROM, RAM, magnetic disk or optical disk and various program code storage media.

[0205] The above description is only specific embodiments of the present application, but the protection scope of the present application is not limited thereto, any person skilled in the art can easily think of changes or replacements within the technical scope disclosed by the present application, which should be covered in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method of detecting a graphic contour, characterized by, The method comprises: acquiring interest surface data and check-in data, each interest surface in the interest surface data containing a corresponding interest surface profile and a geographic location identifier, and the check-in data containing a check-in geographic identifier; calculating a kernel density estimation value of a target position point corresponding to each interest surface in the interest surface data according to a correspondence between the geographic location identifier and the check-in geographic identifier and the interest surface data and the check-in data; generating an isogram of the kernel density estimation value and generating a plurality of isosurfaces according to the isogram; determining a target isosurface corresponding to each interest surface in the interest surface data from the plurality of isosurfaces; detecting an abnormal interest surface profile according to an intersection-union ratio index between the profile of the target isosurface corresponding to each interest surface in the interest surface data and the interest surface profile, comprising: calculating the intersection-union ratio index between the profile of the target isosurface corresponding to each interest surface in the interest surface data and the interest surface profile, specifically: acquiring an intersection operator between the profile of the target isosurface corresponding to each interest surface in the interest surface data and the interest surface profile; acquiring a joint operator between the profile of the target isosurface corresponding to each interest surface in the interest surface data and the interest surface profile; determining a first area operator corresponding to the intersection operator and a second area operator corresponding to the joint operator; and determining the intersection-union ratio index between the profile of the target isosurface corresponding to each interest surface in the interest surface data and the interest surface profile according to a ratio of the first area operator to the second area operator; arranging the intersection-union ratio index from small to large, and sequentially extracting the interest surface profile corresponding to the data in the top k% of the arranged intersection-union ratio index from the interest surface data to determine the abnormal interest surface profile, wherein the k% is an extraction ratio.

2. The graphic profile detection method of claim 1, wherein, The check-in data further contains check-in geographic coordinates, and the calculation of the kernel density estimation value of the target position point corresponding to each interest surface in the interest surface data according to the correspondence between the geographic location identifier and the check-in geographic identifier and the interest surface data and the check-in data comprises: associating the interest surface data and the check-in data according to the correspondence between the geographic location identifier and the check-in geographic identifier to obtain a set of check-in geographic coordinates corresponding to each interest surface in the interest surface data; determining a target position point according to the interest surface profile corresponding to each interest surface in the interest surface data; calculating the kernel density estimation value of the target position point corresponding to each interest surface in the interest surface data according to the set of check-in geographic coordinates.

3. The graphic profile detection method of claim 2, wherein, The calculation of the kernel density estimation value of the target position point corresponding to each interest surface in the interest surface data according to the set of check-in geographic coordinates comprises: determining a target number of check-in geographic coordinates corresponding to each interest surface in the interest surface data according to the set of check-in geographic coordinates; acquiring a target distance from the target position point to the check-in geographic coordinates corresponding to each interest surface in the interest surface data; acquiring a target bandwidth corresponding to a target kernel function; According to the target quantity, the target distance, the target bandwidth and the target kernel function, a kernel density estimation value corresponding to each of the interest surface data is calculated for the target position point, and the kernel density estimation value is used to represent the intensity of the check-in behavior corresponding to the target position point.

4. The graphic profile detection method of claim 3, wherein, The target bandwidth corresponding to the target kernel function is obtained, including: searching for a plurality of candidate bandwidths in a first preset interval according to a preset interval; based on each of the candidate bandwidths, a first kernel density estimation value corresponding to each of the interest surface data is calculated for the target position point; a first contour line of the first kernel density estimation value is generated, and a first contour surface is generated according to the first contour line; a first target contour surface is determined according to the spatial projection area of the first contour surface and the spatial projection area of the contour profile corresponding to the first contour surface; a first intersection and union ratio index between the contour profile of the first target contour surface corresponding to each of the interest surface data and the contour profile of the interest surface is calculated; the highest value in the average value of the first intersection and union ratio index of all interest surfaces in the interest surface data is determined as the candidate bandwidth corresponding to the target kernel function.

5. The graphic profile detection method of claim 2, wherein, The target position point is determined according to the contour profile corresponding to each of the interest surface data, including: obtaining the four-to-point of the contour profile corresponding to each of the interest surface data in the planar coordinate system; generating a rectangular polygon according to the four-to-point to obtain the contour four-to-range of the contour profile corresponding to each of the interest surface data; dotting the contour four-to-range according to a preset grid to obtain a plurality of position points corresponding to each of the contour four-to-range; the target position point is determined from the plurality of position points corresponding to each of the contour four-to-range.

6. The graphic profile detection method of claim 5, wherein, The contour line of the kernel density estimation value is generated, and a plurality of contour surfaces are generated according to the contour line, including: connecting the points with the same value in the kernel density estimation value into a curve according to the contour line generation algorithm to generate a plurality of contour lines, wherein the spatial contour formed by the combination of each of the contour lines and the rectangular boundary of the contour four-to-range is a closed ring line; the closed polygon formed inside the ring line corresponding to each of the contour lines in the plurality of contour lines is taken as a contour surface to obtain a plurality of contour surfaces.

7. The method of claim 1, wherein the step of detecting the graph profile is performed by using a graph profile detection algorithm. The target contour surface corresponding to each of the interest surface data is determined from the plurality of contour surfaces, including: the target contour surface corresponding to each of the interest surface data is determined according to the spatial projection area of the contour surface and the spatial projection area of the contour profile corresponding to the contour surface.

8. The graphic profile detection method of claim 7, wherein, The target contour surface corresponding to each of the interest surface data is determined according to the spatial projection area of the contour surface and the spatial projection area of the contour profile corresponding to the contour surface, including: the spatial projection area of the plurality of contour surfaces and the spatial projection area of the contour profile corresponding to each of the contour surfaces in the plurality of contour surfaces are calculated; Select an isosurface closest to the spatial projection area of the corresponding target interest surface profile from the spatial projection areas of the plurality of isosurfaces, as the target isosurface corresponding to the target interest surface.

9. The method of claim 1, wherein, The method further comprises: sending the abnormal interest surface profile to a target terminal; receiving a cause analysis result of the abnormal interest surface profile fed back by the target terminal; updating the interest surface data according to the cause analysis result.

10. A pattern profile detection apparatus characterized by comprising: The device comprises: an acquisition module configured to acquire interest surface data and check-in data, each interest surface in the interest surface data comprising a corresponding interest surface profile and a geographic location identifier, and the check-in data comprising a check-in geographic identifier; a calculation module configured to calculate a kernel density estimation value of a target location point corresponding to each interest surface in the interest surface data according to a correspondence between the geographic location identifier and the check-in geographic identifier, and the interest surface data and the check-in data; a generation module configured to generate contour lines of the kernel density estimation value, and generate a plurality of isosurfaces according to the contour lines; a determination module configured to determine a target isosurface corresponding to each interest surface in the interest surface data from the plurality of isosurfaces; a detection module configured to detect an abnormal interest surface profile according to an intersection and union ratio index between a profile of the target isosurface corresponding to each interest surface in the interest surface data and the interest surface profile, including: obtaining an intersection operator between the profile of the target isosurface corresponding to each interest surface in the interest surface data and the interest surface profile; obtaining a joint operator between the profile of the target isosurface corresponding to each interest surface in the interest surface data and the interest surface profile; determining a first area operator corresponding to the intersection operator, and a second area operator corresponding to the joint operator; determining the intersection and union ratio index between the profile of the target isosurface corresponding to each interest surface in the interest surface data and the interest surface profile according to a ratio of the first area operator to the second area operator; arranging the intersection and union ratio index from small to large, sequentially extracting interest surface profiles corresponding to top k% data in the arranged intersection and union ratio index from the interest surface data, and determining the interest surface profiles as abnormal interest surface profiles, where the k% is an extraction ratio.

11. A computer readable storage medium, characterized in that, The computer readable storage medium stores a computer program, and the computer program is adapted to be loaded by a processor to execute steps in the graph profile detection method according to any one of claims 1-10.

12. A computer device, comprising: The computer device comprises a processor and a memory, and the memory stores a computer program, and the processor is configured to execute steps in the graph profile detection method according to any one of claims 1-10 by invoking the computer program stored in the memory.

13. A computer program product comprising computer instructions, characterized in that, The computer instructions are executed by the processor to implement steps in the graph profile detection method according to any one of claims 1-10.

Images