Method and device for automatically identifying geofence quality problem
Through the neural network combining the latitude and longitude information of geofences and address scene type relationships, the quality problems of geofences are automatically identified, solving the problem of low manual detection efficiency in the existing technology, and achieving efficient and accurate geofence quality detection.
Patent Information
- Application Number
- CN202411514863.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-29
- Publication Date
- 2025-02-25
AI Technical Summary
In the prior art, the quality inspection of geofences relies on manual operations, is inefficient and time-consuming, making it difficult to efficiently and accurately detect the deviation, range size and shape of the fence.
By obtaining the longitude and latitude information of the geofence, the corresponding map image is determined, and the relationship weights between geofences are determined based on the address scene type, and the relationship matrix data is generated. Enter standard image data and relation matrix data into neural network to automatically identify geofence quality problems.
It realizes the efficiency and accuracy of geofence quality inspection, replaces manual inspection, significantly improves the detection efficiency, and can accurately judge the position deviation, range size and shape of the fence.
Smart Images
Figure CN119521128A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of geographic image processing technology, and in particular to a method and device for automatically identifying quality issues of geographic fences. Background Art
[0002] A geofence is a virtual boundary that represents a geographic area in the real world. In a geographic information system (GIS), a geofence is usually represented by an area of interest (AOI) formed by a closed polygon. A geofence can be dynamically generated by generating a closed surface of a certain range around a point of interest; it can also be generated by matching the area of a pre-defined scene (such as a campus, community, etc.), for example, by using digital image processing methods to extract the border corresponding to the place name from a satellite map and a road network base map.
[0003] Geographical fences play an important role in the digitalization and automation of various industries, such as security and wireless communication planning. If the latitude and longitude of the fence are offset, the fence is too small, or the fence is too large, it is very likely to cause losses to production and life. Therefore, it is necessary to check the quality of the generated geo-fences. At present, the inspection of geo-fences is completed one by one by hand based on experience, but the data of geo-fences is usually in the tens of millions, so it takes a lot of manpower and is inefficient. Summary of the invention
[0004] The present invention provides a method and device for automatically identifying geo-fence quality problems. The method can efficiently and accurately detect the quality of geo-fences. The specific scheme includes: Determine a map image corresponding to the geo-fence according to the acquired fence longitude information and fence latitude information of the geo-fence; Based on the address information of each geo-fence, determine the address scenario type to which each geo-fence belongs; Determine, according to the address scenario type, a relationship weight between each geo-fence, wherein the relationship weight is used to illustrate a subordinate relationship between each geo-fence; generating relationship matrix data based on the relationship weights; The standard image data corresponding to the map image and the relationship matrix data corresponding to the corresponding relationship and the relationship weight are input into the data layer corresponding to the neural network to obtain the quality problem corresponding to the geographic fence.
[0005] Specifically, determining a map image corresponding to the geo-fence according to the acquired fence longitude information and fence latitude information of the geo-fence includes: When the longitude span value and the latitude span value are both greater than the preset target image pixel value: The sum of the maximum longitude value of the geo-fence and the preset offset is used as the maximum longitude value of the map of the map image; The difference between the minimum longitude value of the geo-fence and the preset offset is used as the minimum longitude value of the map of the map image; The sum of the maximum dimension value of the geo-fence and the preset offset is used as the maximum latitude value of the map of the map image; The difference between the minimum latitude value of the geo-fence and the preset offset is used as the minimum latitude value of the map of the map image; The map minimum longitude value, the map maximum longitude value, the map maximum latitude value and the map minimum latitude value are determined as the map longitude value and the map latitude value.
[0006] Specifically, determining the map longitude value and the map latitude value according to the relationship between the longitude span value, the latitude span value and the preset target image pixel value includes: When the longitude span value and the latitude span value are both smaller than the preset target image pixel value, and the pixel value corresponding to the geo-fence is smaller than the preset pixel value: Determine the maximum longitude value and the minimum longitude value of the map image according to the maximum longitude value, the minimum longitude value and the value of the center point in longitude of the geo-fence; Determine the maximum latitude value and the minimum latitude value of the map image according to the maximum latitude value, the minimum latitude value and the latitude value of the center point of the geo-fence; The maximum longitude value of the map, the minimum longitude value of the map, the maximum latitude value of the map, and the minimum latitude value of the map are determined as the map longitude value and the map latitude value.
[0007] Specifically, determining the map longitude value and the map latitude value according to the relationship between the longitude span value, the latitude span value and the preset target image pixel value includes: When the longitude span value and the latitude span value are both smaller than the preset target image pixel value, and the pixel value corresponding to the geo-fence is greater than the preset pixel value: According to the maximum longitude value and the minimum longitude value of the geo-fence, the multiples of the pixel value corresponding to the geo-fence, and the multiples of the longitude span value, respectively determine the maximum longitude value and the minimum longitude value of the map; Determine the maximum latitude value and the minimum latitude value of the map according to the maximum latitude value and the minimum latitude value of the geo-fence, the multiples of the pixel value corresponding to the geo-fence, and the multiples of the latitude span value; The maximum longitude value of the map, the minimum longitude value of the map, the maximum latitude value of the map, and the minimum latitude value of the map are determined as the map longitude value and the map latitude value.
[0008] Specifically, determining the relationship weights between the geo-fences according to the address scenario type includes: According to the level to which the address scenario type belongs, determine the relationship weight between the geo-fences corresponding to the high-level address scenario type and the low-level address scenario type; and the relationship weight between the geo-fences corresponding to the low-level address scenario type.
[0009] Specifically, the vector map and the satellite map corresponding to the geographic fence are respectively input into the convolution layer, and then passed through the MaxPooling and Flatten layers to obtain the standard image data; Processing the address scenario type and the relationship weight using dense layer processing to obtain the relationship matrix data; The standard image data and the relationship matrix data are aggregated together through an aggregation layer to obtain the quality problem.
[0010] The present invention also provides a device for automatically identifying geo-fence quality problems, the device comprising: A first determination module, configured to determine a map image corresponding to the geo-fence according to the acquired fence longitude information and fence latitude information of the geo-fence; A second determination module, configured to determine the address scenario type to which each geo-fence belongs based on the address information of each geo-fence; A third determination module, configured to determine a relationship weight between each geo-fence according to the address scenario type, wherein the relationship weight is used to illustrate a subordinate relationship between each geo-fence; The acquisition module is used to input the standard image data corresponding to the map image and the relationship matrix data corresponding to the corresponding relationship and the relationship weight into the data layer corresponding to the neural network to obtain the quality problem corresponding to the geographic fence.
[0011] The present invention also provides an electronic device, characterized in that the electronic device includes a processor, a memory, and a program or instruction stored in the memory and executable on the processor, wherein the program or instruction is executed by the processor to perform the steps of the above-mentioned method for automatically identifying geo-fence quality issues.
[0012] According to the method for automatically identifying geo-fence quality problems provided by this solution, a map image of appropriate size can be obtained. When the map image is of appropriate size, a geo-fence of appropriate size can be obtained, and such a geo-fence can accurately enclose the target location. Then, the relationship matrix data between the geo-fences is represented by a topological diagram to determine whether the geo-fence belongs to the following situations: fence position offset, fence range too large, fence range too small, fence shape, and fence correct. In this way, the quality of the geo-fence can be accurately judged. BRIEF DESCRIPTION OF THE DRAWINGS
[0013] Figure 1 This is a schematic diagram of a map image; Figure 2 A schematic diagram of a map image with lakes; Figure 3 A flow chart of a method for automatically identifying geo-fence quality issues provided by an embodiment of the present invention; Figure 4 A flow chart of a method for determining a map image; Figure 5 It is a geo-fence relationship diagram; Figure 6 This is a diagram of the neural network framework structure involved in the embodiment of the present application; Figure 7 A structural diagram of a device for automatically identifying geo-fence quality issues provided in an embodiment of the present application. DETAILED DESCRIPTION
[0014] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are described in detail below in conjunction with the accompanying drawings. In the following description, many specific details are set forth to facilitate a full understanding of the present application. However, the present application can be implemented in many other ways different from those described herein, and those skilled in the art can make similar improvements without violating the connotation of the present application, so the present application is not limited by the specific embodiments disclosed below.
[0015] In the description of the present application, it should be understood that if the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise", "axial", "radial", "circumferential", etc. appear, the orientation or position relationship indicated by these terms is based on the orientation or position relationship shown in the accompanying drawings, which is only for the convenience of describing the present application and simplifying the description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present application.
[0016] In addition, if the terms "first" or "second" appear, these terms are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of this application, if the term "plurality" appears, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.
[0017] In this application, unless otherwise clearly specified and limited, if the terms "installed", "connected", "connected", "fixed" and the like appear, these terms should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integrated connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements, unless otherwise clearly defined. For ordinary technicians in this field, the specific meanings of the above terms in this application can be understood according to the specific circumstances.
[0018] In the present application, unless otherwise clearly specified and limited, if there is a description that a first feature is "on" or "below" a second feature, etc., or the like, it may mean that the first and second features are in direct contact, or that the first and second features are in indirect contact through an intermediate medium. Moreover, the first feature being "above", "above" and "above" the second feature may mean that the first feature is directly above or obliquely above the second feature, or simply means that the first feature is higher in level than the second feature. The first feature being "below", "below" and "below" the second feature may mean that the first feature is directly below or obliquely below the second feature, or simply means that the first feature is lower in level than the second feature.
[0019] It should be noted that if an element is referred to as being "fixed to" or "disposed on" another element, it may be directly on the other element or there may be a central element. If an element is considered to be "connected to" another element, it may be directly connected to the other element or there may be a central element at the same time. If any, the terms "vertical", "horizontal", "upper", "lower", "left", "right" and similar expressions used in this application are for illustrative purposes only and do not represent the only implementation method.
[0020] An embodiment of the present application provides a method for automatically identifying geo-fence quality issues. The method is used to automatically identify geo-fence quality by building a machine learning model, replacing the original working method that requires a lot of manpower and is inefficient.
[0021] Geofences can be represented as closed polygons in vector or satellite maps, for example Figure 1The map image shown, where Figure 1 (a) is a vector map containing the geofence, Figure 1 (b) is a satellite map containing the geo-fence, where the closed figure is the geo-fence. Vector maps and satellite maps can provide information such as the latitude and longitude, range, etc. of the geo-fence.
[0022] The information provided by the vector map and satellite map is used as the input data of the machine learning model. The size of these two maps needs to match the location and range of the geofence, otherwise the model will not have enough information to check the quality of the geofence. For example, if the input vector map and satellite map range is too large, it will not only waste computing and storage resources, but also may reduce the accuracy of judgment due to excessive interference from the surrounding environment; if the input map range is too small, the satellite map may fail to load due to insufficient resolution, resulting in missing satellite map data. Figure 2 As shown, suppose you need to obtain the geographic fence of the lake in the figure, the goal is to obtain Figure 2 The left image in the middle, but if the map image obtained is too large ( Figure 2 The right picture in the middle) may result in obtaining an erroneous geo-fence, for example, two adjacent lakes are included in the same geo-fence. Therefore, being able to obtain a suitable map image has a direct impact on the accuracy of the geo-fence. Therefore, inputting a suitable map image has an important impact on the detection of the quality of the geo-fence. Step 31 in this method provides a method for obtaining a suitable map image.
[0023] A method for automatically identifying geo-fence quality issues Figure 3 As shown, the specific process is as follows: Step 31, determining a map image corresponding to the geo-fence according to the acquired fence longitude information and fence latitude information of the geo-fence; This step 31 includes the following processes: Figure 4 As shown: Step 41, reading fence longitude information and fence latitude information of the geographic fence from the vector map and the satellite map corresponding to the map image; The above-mentioned determination of the longitude span value of the geographic fence based on the fence longitude information can be completed by step 42 and step 43.
[0024] Step 42: Obtain the minimum longitude information of the fence from the longitude information of the fence And the maximum longitude information of the fence ; Step 43: Based on the minimum longitude information of the fence And the maximum longitude information of the fence , determining the longitude span value of the geo-fence; The above-mentioned determining the latitude span value of the geo-fence according to the fence latitude information includes step 44 and step 45; Step 44, obtaining the minimum latitude information of the fence from the latitude information of the fence , and the maximum latitude of the fence ; Step 45: Based on the minimum latitude information of the fence , and the maximum latitude of the fence , determining the latitude span value of the geo-fence; Available Represents the conversion function from longitude and latitude to meters, which means the longitude span of the geofence is , indicating that the latitude span of the geofence is ;use and Respectively represent the offsets in longitude and latitude at a specific distance near the geofence, i.e., the preset offsets. For example, The longitude value representing the 5-meter offset around the geofence.
[0025] use Indicates the maximum longitude in the map image, that is, the maximum longitude value of the map; Indicates the minimum longitude value of the map within the map image; Indicates the maximum latitude value of the map within the map image. Indicates the minimum latitude value of the map within the map image.
[0026] According to the relationship between the longitude span value, the latitude span value and the preset target image pixel value, the map longitude value and the map latitude value are determined; the specific process of this step includes one of the following steps 46 to 48: For example, if the standard map image size required by the input model is 128*128, the preset target image pixel value is 128. Satellite maps are processed as electronic grids. The map resolution can be understood as the actual distance represented by a pixel. The satellite map resolution is represented by r. The resolution of satellite maps is between 5m and 50m. The resolution in dense urban areas can reach 5m, and the resolution in rural areas is usually 50m.
[0027] Step 46, when the longitude span value and the latitude span value are both greater than the preset target image pixel value 128, it means that the map resolution is high and can meet the geo-fence range detection requirements. At this time, it is sufficient to ensure that the map range includes the geo-fence.
[0028] According to the map maximum longitude, map minimum longitude, map maximum latitude and map minimum latitude of the geographic image as shown in formula (I), the map minimum longitude value, map maximum longitude value, map maximum latitude value and map minimum latitude value are determined as the map longitude value and map latitude value.
[0029]
[0030]
[0031]
[0032] Formula (I) The above formula (I) is explained as follows: The sum of the maximum longitude value of the geo-fence and the preset offset is used as the maximum longitude value of the map of the map image; The difference between the minimum longitude value of the geo-fence and the preset offset is used as the minimum longitude value of the map of the map image; The sum of the maximum latitude value of the geo-fence and the preset offset is used as the maximum latitude value of the map of the map image; The difference between the minimum latitude value of the geo-fence and the preset offset is used as the minimum latitude value of the map of the map image; Step 47, when the longitude span value and the latitude span value are both less than the preset target image pixel value 128, and the pixel value corresponding to the geo-fence is less than the preset pixel value, for example, r<20m, calculate according to formula (II): , , ,
[0033] Formula (II) The above formula (II) is explained as follows: According to the maximum longitude value, minimum longitude value and center point of the geo-fence The value in longitude is used to determine the maximum longitude value and the minimum longitude value of the map image; Determine the maximum latitude value and the minimum latitude value of the map image according to the maximum latitude value, the minimum latitude value and the latitude value of the center point of the geo-fence; The maximum longitude value of the map, the minimum longitude value of the map, the maximum latitude value of the map, and the minimum latitude value of the map are determined as the map longitude value and the map latitude value.
[0034] Step 48, when the longitude span value and the latitude span value are both less than the preset target image pixel 128, and the pixel value corresponding to the geo-fence is greater than the preset pixel value, for example, r>20m, calculation is performed according to formula (III): , ,
[0035] ,
[0036] Formula (III) The above-mentioned Notice (III) is explained as follows: According to the maximum longitude value, the minimum longitude value, and the multiple of the pixel value corresponding to the geo-fence , multiples of the longitude span value , respectively determine the maximum longitude value and the minimum longitude value of the map; According to the maximum latitude value, the minimum latitude value, and the multiple of the pixel value corresponding to the geo-fence , multiples of the latitude span value , respectively determine the maximum latitude value and the minimum latitude value of the map; The maximum longitude value of the map, the minimum longitude value of the map, the maximum latitude value of the map, and the minimum latitude value of the map are determined as the map longitude value and the map latitude value.
[0037] Formula 3 indicates that the maximum and minimum longitude and latitude of the geographic fence should be expanded by at least 10 pixels. This ensures that the satellite map contains enough pixels. Otherwise, it is easy to cause satellite map missing.
[0038] The map image can be obtained through the above steps 41 to 48, that is, the map image can be drawn according to the map longitude value and the map latitude value.
[0039] Step 32, based on the address information of each geo-fence, determining the address scene type to which each geo-fence belongs; According to the address scene classification, the address scene type can be divided into, for example, school, affiliated middle school, hospital, village, town, etc. This information can be obtained together when obtaining the geographic fence.
[0040] Step 33, according to the address scenario type, determine the relationship weights between the various geo-fences, and the relationship weights are used to illustrate the subordinate relationships between the various geo-fences; in this step, according to the level to which the address scenario type belongs, determine the relationship weights between the geo-fences corresponding to the high-level address scenario type and the low-level address scenario type; the relationship weights between the various geo-fences corresponding to the low-level address scenario type.
[0041] Step 34, generating relationship matrix data based on the relationship weights; Step 35, inputting the standard image data corresponding to the map image, and the relationship matrix data corresponding to the corresponding relationship and the relationship weight into the data layer corresponding to the neural network, and obtaining the quality problem corresponding to the geographic fence.
[0042] For steps 32 to 35, the specific process is described below by way of example: The difference between address scene types and ordinary data types is that there are relationships such as inclusion and subordination between address scene types. If the address scene types to which geofences belong are encoded as ordinary data types, these relationship information will be lost; and compared with natural language, these address scene type data are more regular. If geofences are processed as natural language, a lot of computing power is required for semantic recognition. Therefore, we convert geographic scene types into connection matrices used to express topological graphs, and achieve the encoding of relationships between different geographic scene types with light computing power.
[0043] like Figure 5 As shown in the figure, the geo-fence relationship diagram is used as an example to encode the address scenario type. Figure 5 Medium Represents a topological graph, V represents the nodes in it, and E represents the relationship between nodes.
[0044] Figure 5 The "node" in the figure represents the address scenario type of the geo-fence. There are two types of nodes in the figure, namely the first-level scenario (high-level address scenario type) and the second-level scenario (low-level address scenario type) (orange and blue circles in the figure), denoted as . There are 3, recorded as ,same Recorded as , there are 6 nodes in total.
[0045] Figure 5 The "edge" in the address scene type is the relationship between nodes. For Node Specific relationship It can be written as: We define a weight for each relationship. The weight of express, .when When , it indicates that the node right No relationship ;otherwise , indicating a node right Existence relationship .
[0046] There are two types of relationships in this diagram, namely "include" and "subordinate". The "include" relationship is recorded as , we set the weights ,when When , it indicates the high-level address scenario type Include The “subordinate” relationship is recorded as , we set the weights when When , it indicates the low-level address scenario type Belongs to the primary or secondary address scenario type .
[0047] We use the M-order matrix represents the association matrix, where K is the number of nodes in the topology graph G. In this case, M=6, and the correspondence between nodes and geo-fence scene categories is:
[0048] The value of each element in the incidence matrix , for example, the node (School) and Node (colleges) have a "contains" relationship, then When the scene type of the geofence is "school", the type data input to the machine learning model is two arrays: and .
[0049] In order to allow machine learning to fully extract features from input data, this solution designs a mixed input neural network framework with the following data flow: Figure 6As shown, the satellite map containing geofences, the vector map containing geofences, and the address scene type of geofences are first processed into standard image data and relationship matrix data that can be read by the model according to the above steps. The vector map and satellite map are respectively input into two convolutional layers, and then pass through the MaxPooling and Flatten layers, while the relationship matrix data is directly processed by the dense layer. After the three inputs are aggregated together through the aggregation layer, they pass through the fully connected layer and output quality problems. Quality problems include: fence position offset, fence range too large, fence range too small, fence shape, and fence correctness.
[0050] The machine learning model proposed in this scheme uses the manually inspected geographic fences as the training data set. According to the machine learning model framework passed through the above steps, three input data in the training data set correspond to one output data, and the data format is shown in Table 1.
[0051]
[0052] According to the method for automatically identifying geo-fence quality problems provided by this solution, a map image of appropriate size can be obtained. When the map image is of appropriate size, a geo-fence of appropriate size can be obtained, and such a geo-fence can accurately enclose the target location. Then, the relationship matrix data between the geo-fences is represented by a topological diagram to determine whether the geo-fence belongs to the following situations: fence position offset, fence range too large, fence range too small, fence shape, and fence correct. In this way, the quality of the geo-fence can be accurately judged.
[0053] The present invention provides a device for automatically identifying geo-fence quality problems, such as Figure 7 As shown, the device comprises: A first determination module 71, configured to determine a map image corresponding to the geo-fence according to the acquired fence longitude information and fence latitude information of the geo-fence; A second determination module 72, configured to determine the address scenario type to which each geo-fence belongs based on the address information of each geo-fence; A third determination module 73, configured to determine a relationship weight between each geo-fence according to the address scenario type, wherein the relationship weight is used to illustrate a subordinate relationship between each geo-fence; A generating module 74, configured to generate relationship matrix data based on the relationship weights; The acquisition module 75 is used to input the standard image data corresponding to the map image and the relationship matrix data corresponding to the corresponding relationship and the relationship weight into the data layer corresponding to the neural network to obtain the quality problem corresponding to the geographic fence.
[0054] The first determination module 71 is specifically used to determine the longitude span value of the geographic fence according to the fence longitude information; determine the latitude span value of the geographic fence according to the fence latitude information; determine the map longitude value and the map latitude value according to the relationship between the longitude span value, the latitude span value and the preset target image pixel value; the range determined by the map longitude value and the map latitude value is the map image.
[0055] Wherein, determining the map longitude value and the map latitude value according to the relationship between the longitude span value, the latitude span value and the preset target image pixel value includes: When the longitude span value and the latitude span value are both greater than the preset target image pixel value: the sum of the maximum longitude value of the geofence and the preset offset is used as the map maximum longitude value of the map image; the difference between the minimum longitude value of the geofence and the preset offset is used as the map minimum longitude value of the map image; the sum of the maximum dimension value of the geofence and the preset offset is used as the map maximum latitude value of the map image; the difference between the minimum latitude value of the geofence and the preset offset is used as the map minimum latitude value of the map image; the map minimum longitude value, map maximum longitude value, map maximum latitude value and map minimum latitude value are determined as the map longitude value and map latitude value.
[0056] Wherein, determining the map longitude value and the map latitude value according to the relationship between the longitude span value, the latitude span value and the preset target image pixel value includes: When the longitude span value and the latitude span value are both smaller than the preset target image pixel value, and the pixel value corresponding to the geo-fence is smaller than the preset pixel value: determine the map maximum longitude value and the map minimum longitude value of the map image according to the maximum longitude value, the minimum longitude value and the value of the center point in longitude of the geo-fence; determine the map maximum latitude value and the map minimum latitude value of the map image according to the maximum latitude value, the minimum latitude value and the value of the center point in latitude of the geo-fence; determine the map maximum longitude value, the map minimum longitude value, the map maximum latitude value and the map minimum latitude value as the map longitude value and the map latitude value.
[0057] Wherein, determining the map longitude value and the map latitude value according to the relationship between the longitude span value, the latitude span value and the preset target image pixel value includes: When the longitude span value and the latitude span value are both smaller than the preset target image pixel value, and the pixel value corresponding to the geo-fence is greater than the preset pixel value: determine the maximum longitude value and the minimum longitude value of the map according to the maximum longitude value and the minimum longitude value of the geo-fence, multiples of the pixel value corresponding to the geo-fence, and multiples of the longitude span value; determine the maximum latitude value and the minimum latitude value of the map according to the maximum latitude value and the minimum latitude value of the geo-fence, multiples of the pixel value corresponding to the geo-fence, and multiples of the latitude span value; determine the maximum longitude value, the minimum longitude value, the maximum latitude value and the minimum latitude value of the map as the map longitude value and the map latitude value.
[0058] The third determination module 73 is specifically used to determine, according to the level to which the address scenario type belongs, the relationship weight between the geo-fences corresponding to the high-level address scenario type and the low-level address scenario type; and the relationship weight between the geo-fences corresponding to the low-level address scenario type.
[0059] The specific device provided by this solution includes the following functions: The vector map and satellite map corresponding to the geographic fence are respectively input into the convolution layer, and then passed through the MaxPooling and Flatten layers to obtain the standard image data; the address scene type and the relationship weight are processed using the dense layer to obtain the relationship matrix data; the standard image data and the relationship matrix data are aggregated together through the aggregation layer to obtain the quality problem.
[0060] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0061] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be construed as limiting the scope of the patent application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent application shall be subject to the attached claims.
Claims
1. A method for automatically identifying geo-fence quality issues, characterized in that: The method comprises: Determine a map image corresponding to the geo-fence according to the acquired fence longitude information and fence latitude information of the geo-fence; Based on the address information of each geo-fence, determine the address scenario type to which each geo-fence belongs; Determine, according to the address scenario type, a relationship weight between each geo-fence, wherein the relationship weight is used to illustrate a subordinate relationship between each geo-fence; generating relationship matrix data based on the relationship weights; The standard image data corresponding to the map image and the relationship matrix data corresponding to the corresponding relationship and the relationship weight are input into the data layer corresponding to the neural network to obtain the quality problem corresponding to the geographic fence.
2. The method according to claim 1, characterized in that The determining a map image corresponding to the geo-fence according to the acquired fence longitude information and fence latitude information of the geo-fence includes: Determine the longitude span value of the geo-fence according to the fence longitude information; Determining a latitude span value of the geo-fence according to the fence latitude information; Determine a map longitude value and a map latitude value according to a relationship between the longitude span value, the latitude span value and a preset target image pixel value; The range determined by the map longitude value and the map latitude value is the map image.
3. The method according to claim 2, characterized in that The determining of the map longitude value and the map latitude value according to the relationship between the longitude span value, the latitude span value and the preset target image pixel value includes: When the longitude span value and the latitude span value are both greater than the preset target image pixel value: The sum of the maximum longitude value of the geo-fence and the preset offset is used as the maximum longitude value of the map of the map image; The difference between the minimum longitude value of the geo-fence and the preset offset is used as the minimum longitude value of the map of the map image; The sum of the maximum dimension value of the geo-fence and the preset offset is used as the maximum latitude value of the map of the map image; The difference between the minimum latitude value of the geo-fence and the preset offset is used as the minimum latitude value of the map of the map image; The map minimum longitude value, the map maximum longitude value, the map maximum latitude value and the map minimum latitude value are determined as the map longitude value and the map latitude value.
4. The method according to claim 2, characterized in that The determining of the map longitude value and the map latitude value according to the relationship between the longitude span value, the latitude span value and the preset target image pixel value includes: When the longitude span value and the latitude span value are both smaller than the preset target image pixel value, and the pixel value corresponding to the geo-fence is smaller than the preset pixel value: Determine the maximum longitude value and the minimum longitude value of the map image according to the maximum longitude value, the minimum longitude value and the value of the center point in longitude of the geo-fence; Determine the maximum latitude value and the minimum latitude value of the map image according to the maximum latitude value, the minimum latitude value and the latitude value of the center point of the geo-fence; The maximum longitude value of the map, the minimum longitude value of the map, the maximum latitude value of the map, and the minimum latitude value of the map are determined as the map longitude value and the map latitude value.
5. The method according to claim 2, characterized in that The determining of the map longitude value and the map latitude value according to the relationship between the longitude span value, the latitude span value and the preset target image pixel value includes: When the longitude span value and the latitude span value are both smaller than the preset target image pixel value, and the pixel value corresponding to the geo-fence is greater than the preset pixel value: According to the maximum longitude value and the minimum longitude value of the geo-fence, the multiples of the pixel value corresponding to the geo-fence, and the multiples of the longitude span value, respectively determine the maximum longitude value and the minimum longitude value of the map; Determine the maximum latitude value and the minimum latitude value of the map according to the maximum latitude value and the minimum latitude value of the geo-fence, the multiples of the pixel value corresponding to the geo-fence, and the multiples of the latitude span value; The maximum longitude value of the map, the minimum longitude value of the map, the maximum latitude value of the map, and the minimum latitude value of the map are determined as the map longitude value and the map latitude value.
6. The method according to claim 1, characterized in that Determining the relationship weights between the geo-fences according to the address scenario type includes: According to the level to which the address scenario type belongs, determine the relationship weight between the geo-fences corresponding to the high-level address scenario type and the low-level address scenario type; and the relationship weight between the geo-fences corresponding to the low-level address scenario type.
7. The method according to claim 1, characterized in that The vector map and satellite map corresponding to the geographic fence are input into the convolution layer respectively, and then passed through the MaxPooling and Flatten layers to obtain the standard image data; Processing the address scenario type and the relationship weight using dense layer processing to obtain the relationship matrix data; The standard image data and the relationship matrix data are aggregated together through an aggregation layer to obtain the quality problem.
8. A device for automatically identifying geo-fence quality issues, characterized in that: The device comprises: A first determination module, configured to determine a map image corresponding to the geo-fence according to the acquired fence longitude information and fence latitude information of the geo-fence; A second determination module, configured to determine the address scenario type to which each geo-fence belongs based on the address information of each geo-fence; A third determination module, configured to determine a relationship weight between each geo-fence according to the address scenario type, wherein the relationship weight is used to illustrate a subordinate relationship between each geo-fence; A generating module, used for generating relationship matrix data based on the relationship weights; The acquisition module is used to input the standard image data corresponding to the map image and the relationship matrix data corresponding to the corresponding relationship and the relationship weight into the data layer corresponding to the neural network to obtain the quality problem corresponding to the geographic fence.
9. An electronic device, characterized in that: The electronic device includes a processor, a memory, and a program or instruction stored in the memory and executable on the processor, wherein the program or instruction is executed by the processor to perform the steps of any method of automatically identifying geo-fence quality issues described in any one of claims 1-7.