Data processing method, prediction model training method, electronic device, and program product

Abstract

The embodiments of the present disclosure disclose a data processing method, a prediction model training method, an electronic device and a program product. The data processing method comprises: acquiring position information of a point of interest; acquiring position information of a specified object located within a first preset range around the point of interest and an image covering geographical elements within the point of interest and a second preset range around the point of interest according to the position information of the point of interest; and using a trained prediction model to predict whether the position information of the point of interest is accurate based on the position information of the point of interest, the position information of the specified object and the image.

Description

Technical Field

[0001] This disclosure relates to the field of data processing technology, specifically to a data processing method, a prediction model training method, an electronic device, and a program product. Background Technology

[0002] In the field of electronic mapping, Points of Interest (POIs) refer to real-world entities such as companies and restaurants, and are data that users pay close attention to when using electronic maps. Because the source of POI location information is complex, its location may contain errors, affecting the accuracy of applications such as route planning and navigation. Therefore, identifying potentially erroneous POI location information from massive amounts of map data is crucial for improving the accuracy of electronic maps and related services. Summary of the Invention

[0003] To address the problems in the related technologies, embodiments of this disclosure provide a data processing method, a prediction model training method, an electronic device, and a program product.

[0004] Firstly, this disclosure provides a data processing method.

[0005] Specifically, the data processing method includes:

[0006] Obtain the location information of points of interest;

[0007] Based on the location information of the point of interest, obtain the location information of a specified object located within a first preset range around the point of interest and an image of ground features covering the point of interest and its surrounding second preset range;

[0008] Based on the location information of the point of interest, the location information of the specified object, and the image, a trained prediction model is used to predict whether the location information of the point of interest is accurate.

[0009] In conjunction with the first aspect, in a first implementation of the first aspect of this disclosure, the location information of the designated object located within a first preset range around the point of interest includes at least one of the following:

[0010] Location information of buildings within a first preset range around the point of interest;

[0011] Location information of roads within a first preset range around the point of interest;

[0012] Location information of water systems within a first preset range around the point of interest;

[0013] Location information of other points of interest located within a first preset range around the point of interest;

[0014] Location information of objects associated with the point of interest within a first preset range around the point of interest.

[0015] In conjunction with the first implementation of the first aspect, in the second implementation of the first aspect, the present disclosure includes:

[0016] The location information of the building includes the location information of multiple sampling points around the horizontal outline of the building;

[0017] The location information of the road includes the location information of multiple sampling points along the road;

[0018] The location information of the water system includes the location information of multiple sampling points along the water system;

[0019] The location information of the object associated with the point of interest includes any one of the following: location information of the navigation destination, location information of the device accessing the local network of the point of interest, and location information of the address associated with the point of interest.

[0020] In conjunction with the first aspect, in the third implementation of the first aspect, the step of using a trained prediction model to predict whether the location information of the interest point is accurate based on the location information of the interest point, the location information of the specified object, and the image includes:

[0021] Based on the location information of the specified object and the location information of the point of interest, extract geographic location features to characterize the relative positional relationship between the specified object and the point of interest;

[0022] From the image, extract image location features that characterize the relative positional relationship between the point of interest and its surrounding features;

[0023] The geographic location features and image location features are input into the trained prediction model to predict whether the location information of the point of interest is accurate.

[0024] In conjunction with the first aspect, in a fourth implementation of the first aspect, the method further includes:

[0025] Acquire the location information of the sample interest points and the information on whether the location information of the sample interest points is accurate;

[0026] Based on the location information of the sample point of interest, obtain the location information of the specified sample object within a first preset range around the sample point of interest and the sample image with a shooting range covering the sample point of interest and its surrounding second preset range;

[0027] The prediction model is trained based on the location information of the sample interest points, the location information of the specified object in the sample, the sample image, and the discrimination information on whether the location information of the sample interest points is accurate.

[0028] Secondly, this disclosure provides a prediction model training method, including:

[0029] Acquire the location information of the sample interest points and the information on whether the location information of the sample interest points is accurate;

[0030] Based on the location information of the sample point of interest, obtain the location information of the specified sample object within a first preset range around the sample point of interest and the sample image with a shooting range covering the sample point of interest and its surrounding second preset range;

[0031] The prediction model is trained based on the location information of the sample interest points, the location information of the specified object in the sample, the sample image, and the discrimination information on whether the location information of the sample interest points is accurate.

[0032] In conjunction with the second aspect, in the first implementation of the second aspect of this disclosure, the location information of the specified object of the sample located within a first preset range around the sample point of interest includes at least one of the following:

[0033] Location information of buildings within a first preset range surrounding the sample point of interest;

[0034] Location information of roads within a first preset range around the sample point of interest;

[0035] Location information of water systems within a first preset range around the sample point of interest;

[0036] Location information of other points of interest located within a first preset range around the sample point of interest;

[0037] Location information of objects associated with the sample point of interest within a first preset range around the sample point of interest.

[0038] In conjunction with the second aspect, in a second implementation of the second aspect of this disclosure, training the prediction model based on the location information of the sample interest points, the location information of the specified sample object, the sample image, and the discrimination information regarding the accuracy of the location information of the sample interest points includes:

[0039] Based on the location information of the specified object in the sample and the location information of the point of interest in the sample, extract sample geographic location features to characterize the relative positional relationship between the specified object in the sample and the point of interest in the sample of this type;

[0040] From the sample images, extract sample image location features to characterize the relative positional relationship between the sample point of interest and its surrounding land features;

[0041] The geographical location features and image location features of the samples are input into the prediction model, and the prediction model is trained with the goal of minimizing the loss function between the output of the prediction model and the discrimination information.

[0042] Thirdly, this disclosure provides a data processing apparatus.

[0043] Specifically, the data processing apparatus includes:

[0044] The first acquisition module is configured to acquire the location information of points of interest.

[0045] The second acquisition module is configured to acquire, based on the location information of the point of interest, the location information of a specified object located within a first preset range around the point of interest and an image of ground features covering the point of interest and its surrounding second preset range.

[0046] The prediction module is configured to use a trained prediction model to predict whether the location information of the point of interest is accurate, based on the location information of the point of interest, the location information of the specified object, and the image.

[0047] In conjunction with the third aspect, in the first implementation of the third aspect of this disclosure, the location information of the designated object located within a first preset range around the point of interest includes at least one of the following:

[0048] Location information of buildings within a first preset range around the point of interest;

[0049] Location information of roads within a first preset range around the point of interest;

[0050] Location information of water systems within a first preset range around the point of interest;

[0051] Location information of other points of interest located within a first preset range around the point of interest;

[0052] Location information of objects associated with the point of interest within a first preset range around the point of interest.

[0053] In conjunction with the first implementation of the third aspect, this disclosure provides a second implementation of the third aspect, wherein:

[0054] The location information of the building includes the location information of multiple sampling points around the horizontal outline of the building;

[0055] The location information of the road includes the location information of multiple sampling points along the road;

[0056] The location information of the water system includes the location information of multiple sampling points along the water system;

[0057] The location information of the object associated with the point of interest includes any one of the following: location information of the navigation destination, location information of the device accessing the local network of the point of interest, and location information of the address associated with the point of interest.

[0058] In conjunction with the third aspect, in a third implementation of the third aspect of this disclosure, the step of using a trained prediction model to predict whether the location information of the interest point is accurate based on the location information of the interest point, the location information of the specified object, and the image includes:

[0059] Based on the location information of the specified object and the location information of the point of interest, extract geographic location features to characterize the relative positional relationship between the specified object and the point of interest;

[0060] From the image, extract image location features that characterize the relative positional relationship between the point of interest and its surrounding features;

[0061] The geographic location features and image location features are input into the trained prediction model to predict whether the location information of the point of interest is accurate.

[0062] In conjunction with the third aspect, in a fourth implementation of the third aspect, the apparatus further includes:

[0063] The third acquisition module is configured to acquire the location information of the sample interest points and the discrimination information on whether the location information of the sample interest points is accurate.

[0064] The fourth acquisition module is configured to acquire, based on the location information of the sample interest point, the location information of the specified sample object within a first preset range around the sample interest point and the image of the sample covering the sample interest point and its surrounding second preset range.

[0065] The training module is configured to train the prediction model based on the location information of the sample interest points, the location information of the specified object in the sample, the sample image, and the discrimination information on whether the location information of the sample interest points is accurate.

[0066] Fourthly, this disclosure provides a prediction model training apparatus, comprising:

[0067] The third acquisition module is configured to acquire the location information of the sample interest points and the discrimination information on whether the location information of the sample interest points is accurate.

[0068] The fourth acquisition module is configured to acquire, based on the location information of the sample interest point, the location information of the specified sample object within a first preset range around the sample interest point and the image of the sample covering the sample interest point and its surrounding second preset range.

[0069] The training module is configured to train the prediction model based on the location information of the sample interest points, the location information of the specified object in the sample, the sample image, and the discrimination information on whether the location information of the sample interest points is accurate.

[0070] In conjunction with the fourth aspect, in the first implementation of the fourth aspect of this disclosure, the location information of the specified object of the sample located within a first preset range around the sample point of interest includes at least one of the following:

[0071] Location information of buildings within a first preset range surrounding the sample point of interest;

[0072] Location information of roads within a first preset range around the sample point of interest;

[0073] Location information of water systems within a first preset range around the sample point of interest;

[0074] Location information of other points of interest located within a first preset range around the sample point of interest;

[0075] Location information of objects associated with the sample point of interest within a first preset range around the sample point of interest.

[0076] In conjunction with the fourth aspect, in the second implementation of the fourth aspect of this disclosure, training the prediction model based on the location information of the sample interest points, the location information of the specified object in the sample, the sample image, and the discrimination information regarding the accuracy of the location information of the sample interest points includes:

[0077] Based on the location information of the specified object in the sample and the location information of the point of interest in the sample, extract sample geographic location features to characterize the relative positional relationship between the specified object in the sample and the point of interest in the sample of this type;

[0078] From the sample images, extract sample image location features to characterize the relative positional relationship between the sample point of interest and its surrounding land features;

[0079] The geographical location features and image location features of the samples are input into the prediction model, and the prediction model is trained with the goal of minimizing the loss function between the output of the prediction model and the discrimination information.

[0080] Fifthly, embodiments of this disclosure provide an electronic device including a memory and a processor, wherein the memory is used to store one or more computer instructions, wherein the one or more computer instructions are executed by the processor to implement the method as described in any one of the first to second aspects, the second implementation.

[0081] In a sixth aspect, this disclosure provides a computer-readable storage medium having computer instructions stored thereon, which, when executed by a processor, implement the method described in the second implementation of the first to second aspects.

[0082] In a seventh aspect, this disclosure provides a computer program product including computer instructions that, when executed by a processor, implement the method described in the second implementation of the first to second aspects.

[0083] According to the technical solution provided in the embodiments of this disclosure, location information of points of interest (POIs) is obtained; based on the location information of the POIs, location information of designated objects within a first preset range around the POIs and images of ground features within a second preset range around the POIs are obtained; based on the location information of the POIs, the location information of the designated objects, and the images, a trained prediction model is used to predict whether the location information of the POIs is accurate. According to the embodiments of this disclosure, the prediction model is trained based on the location information of sample POIs, the location information of sample designated objects within a first preset range around the sample POIs, and sample images of ground features within a second preset range around the sample POIs. The prediction model learns the judgment rules, eliminating the need for users to possess strong domain knowledge to create features. Furthermore, the data input to the prediction model suffers less accuracy loss compared to the original data, thus solving the problem of accuracy degradation during feature creation.

[0084] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0085] Other features, objects, and advantages of this disclosure will become more apparent from the following detailed description of non-limiting embodiments, taken in conjunction with the accompanying drawings. In the drawings:

[0086] Figure 1 A flowchart illustrating a data processing method according to an embodiment of the present disclosure is shown.

[0087] Figures 2A-2F A schematic diagram illustrating the relative positional relationship between a designated object and a point of interest according to an embodiment of this disclosure is shown.

[0088] Figure 3 A flowchart illustrating a predictive model training method according to an embodiment of the present disclosure is shown.

[0089] Figure 4 A schematic flowchart of a data processing method according to an embodiment of the present disclosure is shown.

[0090] Figure 5A A structural block diagram of a data processing apparatus according to an embodiment of the present disclosure is shown.

[0091] Figure 5B A structural block diagram of a predictive model training apparatus according to an embodiment of the present disclosure is shown.

[0092] Figure 6 A structural block diagram of an electronic device according to an embodiment of the present disclosure is shown.

[0093] Figure 7 A schematic diagram of the structure of a computer system suitable for implementing the method according to embodiments of the present disclosure is shown. Detailed Implementation

[0094] In the following, exemplary embodiments of the present disclosure will be described in detail with reference to the accompanying drawings to enable those skilled in the art to readily implement them. Furthermore, for clarity, portions unrelated to the description of exemplary embodiments have been omitted from the drawings.

[0095] In this disclosure, it should be understood that terms such as “comprising” or “having” are intended to indicate the presence of features, figures, steps, behaviors, components, parts or combinations thereof disclosed in this specification, and are not intended to exclude the possibility of the presence or addition of one or more other features, figures, steps, behaviors, components, parts or combinations thereof.

[0096] It should also be noted that, unless otherwise specified, the embodiments and features described in this disclosure can be combined with each other. This disclosure will now be described in detail with reference to the accompanying drawings and embodiments.

[0097] In this disclosure, the acquisition of user information or user data is done with the user's authorization or confirmation, or by the user's active choice.

[0098] As mentioned above, due to the complex sources of POI location information, its location may contain errors, affecting the accuracy of applications such as route planning and navigation. Identifying potentially erroneous POI location information from massive amounts of map data is of great significance for improving the accuracy of electronic maps.

[0099] This disclosure provides a data processing method, including: acquiring location information of a point of interest; acquiring location information of a designated object located within a first preset range around the point of interest and an image of ground features covering the point of interest and its surrounding second preset range, based on the location information of the point of interest, the location information of the designated object, and the image; and using a trained prediction model to predict whether the location information of the point of interest is accurate, based on the location information of the point of interest, the location information of the designated object, and the image.

[0100] Existing methods for determining the accuracy of Points of Interest (POI) location information include summarizing judgment rules based on map domain knowledge, such as POIs not being located in water or on roads, to assess the accuracy of POI location information. Alternatively, machine learning can be used, employing coordinate clustering and spatial topology algorithms to generate features, and then using machine learning algorithms such as SVM (Support Vector Machine) and GBDT (Gradient Boosting Decision Tree) to discover erroneous POI location information. These methods have the following problems: First, feature generation is difficult; creating complete features requires strong domain knowledge from the feature extractor. Second, feature generation requires processing of the raw data, including but not limited to aggregation and fusion, which reduces data accuracy.

[0101] According to embodiments of this disclosure, the prediction model is trained on the location information of sample points of interest, the location information of sample designated objects within a first preset range around the sample points of interest, and sample images of ground features within a second preset range covering the sample points of interest and their surrounding areas. The prediction model learns judgment rules, eliminating the need for users to have strong domain knowledge to create features. At the same time, the data input to the prediction model has less accuracy loss compared to the original data, thus solving the problem of accuracy degradation during feature creation.

[0102] Figure 1 A flowchart illustrating a data processing method according to an embodiment of the present disclosure is shown. Figure 1 As shown, the data processing method includes the following steps S101-S103:

[0103] In step S101, the location information of the point of interest is obtained;

[0104] In step S102, based on the location information of the point of interest, the location information of a specified object located within a first preset range around the point of interest and an image of ground features covering the point of interest and its surrounding second preset range are obtained.

[0105] In step S103, based on the location information of the point of interest, the location information of the specified object, and the image, a trained prediction model is used to predict whether the location information of the point of interest is accurate.

[0106] According to embodiments of this disclosure, the location information of the point of interest and the location information of the designated object can be obtained from a map database and / or a location database. The image covering the point of interest and its surrounding features within a second preset range can be a satellite image taken by a satellite in the air. Satellite images can capture ground features, but their clarity and resolution may be low. Alternatively, the image can also be an image taken on the ground around the point of interest, which can clearly capture the point of interest and its surrounding features.

[0107] According to embodiments of this disclosure, the first preset range may be the same as or different from the second preset range.

[0108] Geographical features are a general term for all tangible objects (such as mountains, forests, water systems, roads, buildings, etc.) and intangible objects (such as provincial and county boundaries) on the Earth's surface, referring to relatively fixed objects on the Earth's surface.

[0109] According to embodiments of this disclosure, the prediction model may include, for example, any one or more convolutional neural networks such as LeNet, AlexNet, and VGGNet, or may include an improved ResNet neural network. The improved ResNet neural network omits the third and fourth stages of the ResNet neural network, directly connecting the output of the second stage to the output stage.

[0110] According to embodiments of this disclosure, the prediction model is trained on the location information of sample points of interest, the location information of sample designated objects within a first preset range around the sample points of interest, and sample images of ground features within a second preset range covering the sample points of interest and their surrounding areas. The prediction model learns judgment rules, eliminating the need for users to have strong domain knowledge to create features. At the same time, the data input to the prediction model has less accuracy loss compared to the original data, thus solving the problem of accuracy degradation during feature creation.

[0111] According to embodiments of this disclosure, the location information of the designated object located within a first preset range around the point of interest includes at least one of the following:

[0112] Location information of buildings within a first preset range around the point of interest;

[0113] Location information of roads within a first preset range around the point of interest;

[0114] Location information of water systems within a first preset range around the point of interest;

[0115] Location information of other points of interest located within a first preset range around the point of interest;

[0116] Location information of objects associated with the point of interest within a first preset range around the point of interest.

[0117] According to embodiments of this disclosure, location information may include, for example, coordinate information, such as latitude and longitude.

[0118] According to embodiments of this disclosure, the location information of the building includes location information of multiple sampling points surrounding the horizontal contour of the building;

[0119] The location information of the road includes the location information of multiple sampling points along the road;

[0120] The location information of the water system includes the location information of multiple sampling points along the water system;

[0121] The location information of the object associated with the point of interest includes any one of the following: the location of the navigation destination, the location of the device connected to the local network of the point of interest, and the address associated with the point of interest.

[0122] According to embodiments of this disclosure, points of interest should generally be located inside a building; if a point of interest is located outside a building, its location may be incorrect.

[0123] According to embodiments of this disclosure, points of interest should generally not be located on roads; if a point of interest is located on a road, its location may be incorrect.

[0124] According to embodiments of this disclosure, points of interest should generally not be located on water; if a point of interest is located on water, its location may be incorrect.

[0125] According to embodiments of this disclosure, the distribution pattern of points of interest should generally be consistent with that of other points of interest. If the distribution pattern of points of interest is inconsistent with that of other points of interest, the location of the points of interest may be incorrect.

[0126] According to embodiments of this disclosure, the location information of objects associated with a point of interest (POI) should have a reasonable relationship with the POI's location. For example, if the navigation destination is the POI, the location of the navigation destination should be at or near the POI. The device location of a device accessing the POI's local network should be at or near the POI. The location of the address associated with the POI should be at or near the POI. If the relationship between the location information of objects associated with the POI and the POI's location is unreasonable, the POI's location may be incorrect.

[0127] According to embodiments of this disclosure, the step of predicting the accuracy of the location information of the interest point using a trained prediction model based on the location information of the interest point, the location information of the specified object, and the image includes:

[0128] Based on the location information of the specified object and the location information of the point of interest, extract geographic location features to characterize the relative positional relationship between the specified object and the point of interest;

[0129] From the image, extract image location features that characterize the relative positional relationship between the point of interest and its surrounding features;

[0130] The geographic location features and image location features are input into the trained prediction model to predict whether the location information of the point of interest is accurate.

[0131] Figures 2A-2F A schematic diagram illustrating the relative positional relationship between a designated object and a point of interest according to an embodiment of this disclosure is shown.

[0132] exist Figures 2A-2F In this context, dots indicate the location of points of interest, squares indicate the location of specified objects, and dashed lines and text are used for supplementary explanations.

[0133] like Figures 2A-2F As shown, for various types of specified objects, the relative positional relationship between the specified object and the point of interest is determined based on the positional information of the specified object and the positional information of the point of interest.

[0134] exist Figure 2A In the image, for a specified object "building," there are buildings A and B within a first preset range around the point of interest. The image shows multiple sampling points along the horizontal contours of buildings A and B. As can be seen from the image, the point of interest is located outside of buildings A and B; therefore, the location of the point of interest may be incorrect.

[0135] exist Figure 2BIn the image, for a specified object "road," there is a road C within a first preset range around the point of interest. The figure shows multiple sampling points along road C. As can be seen from the image, the point of interest is located within road C; therefore, the location of this point of interest may be incorrect.

[0136] exist Figure 2C In the image, for a specified object "river" (belonging to the "water system" type), there is a river D within a first preset range around the point of interest. The image shows multiple sampling points along river D. As can be seen from the image, the point of interest is located within river D; therefore, the location of this point of interest may be incorrect.

[0137] exist Figure 2D In the diagram, for a specified object "Other Points of Interest," there are other points of interest a to h distributed along the periphery of building E within a first preset range around the point of interest. As can be seen from the diagram, the distribution of this point of interest differs from the other points of interest; it is located in the center of building E. Therefore, the location of this point of interest may be incorrect.

[0138] exist Figure 2E In the context of the specified object "the object associated with the point of interest", both the object's location and the point of interest are within building F. Therefore, the location of the point of interest is likely correct.

[0139] exist Figure 2F In the context of the specified object "the object associated with the point of interest", the object's location and the point of interest are in building H and building G respectively. Therefore, the location of the point of interest may be incorrect.

[0140] According to embodiments of this disclosure, image location features are extracted from the image to characterize the relative positional relationship between the point of interest and its surrounding features. For example, the position of features in the image can be determined based on the image's shooting range and scale, and then the relative positional relationship between the point of interest and the surrounding features can be obtained.

[0141] The geographic location features and image location features are input into the trained prediction model, and the prediction model outputs a judgment result on whether the location of the point of interest is accurate. Since the location information of the specified object may also be incorrect, and the location relationship between the specified object and the point of interest in the training samples may not cover all possible situations in reality, the image location features extracted from the image within a second preset range around the point of interest are also input into the prediction model. This can complement the location information of the specified object within a first preset range around the point of interest, thereby obtaining a more accurate judgment result.

[0142] Figure 3 A flowchart illustrating a predictive model training method according to embodiments of the present disclosure is shown. Figure 3As shown, the prediction model training method includes the following steps S301–S303:

[0143] In step S301, the location information of the sample interest points and the discrimination information on whether the location information of the sample interest points is accurate are obtained;

[0144] In step S302, based on the location information of the sample point of interest, the location information of the specified sample object located within a first preset range around the sample point of interest and the sample image with a shooting range covering the sample point of interest and its surrounding second preset range are obtained.

[0145] In step S303, the prediction model is trained based on the location information of the sample interest points, the location information of the specified object in the sample, the sample image, and the discrimination information on whether the location information of the sample interest points is accurate.

[0146] According to embodiments of this disclosure, the prediction model is trained based on the location information of the sample interest points, the location information of the specified object of the sample, the sample image, and the discrimination information on whether the location information of the sample interest points is accurate. The prediction model learns the judgment rules, which does not require the user to have strong domain knowledge to create features. At the same time, the data input to the prediction model has less accuracy loss compared to the original data, thus solving the problem of accuracy decline in the feature creation process.

[0147] According to embodiments of this disclosure, the location information of the designated object of the sample located within a first preset range around the sample point of interest includes at least one of the following:

[0148] Location information of buildings within a first preset range surrounding the sample point of interest;

[0149] Location information of roads within a first preset range around the sample point of interest;

[0150] Location information of water systems within a first preset range around the sample point of interest;

[0151] Location information of other points of interest located within a first preset range around the sample point of interest;

[0152] Location information of objects associated with the sample point of interest within a first preset range around the sample point of interest.

[0153] According to embodiments of this disclosure, training the prediction model based on the location information of the sample interest points, the location information of the specified object in the sample, the sample image, and the discrimination information regarding the accuracy of the location information of the sample interest points includes:

[0154] Based on the location information of the specified object in the sample and the location information of the point of interest in the sample, extract sample geographic location features to characterize the relative positional relationship between the specified object in the sample and the point of interest in the sample of this type;

[0155] From the sample images, extract sample image location features to characterize the relative positional relationship between the sample point of interest and its surrounding land features;

[0156] The geographical location features and image location features of the samples are input into the prediction model, and the prediction model is trained with the goal of minimizing the loss function between the output of the prediction model and the discrimination information.

[0157] According to embodiments of this disclosure, the sample data may include data of sample points of interest in correct locations and data of sample points of interest in incorrect locations. The data of sample points of interest in incorrect locations may come from erroneous data reported by users during use and verified by operators, and may also include erroneous data discovered by electronic map service providers in map-making operations such as hotel verification and restaurant verification.

[0158] Figure 4 A schematic flowchart of a data processing method according to an embodiment of the present disclosure is shown.

[0159] like Figure 4 As shown, the sample sources for the prediction model may include, for example, erroneous interest points reported by users and erroneous interest points discovered in map-making operations, as well as a portion of correct sample interest points.

[0160] The system retrieves the location information of sample points of interest (POIs) and the location information of designated sample objects within a first preset range surrounding the POIs from the database, and generates sample geographic location features representing the relative positional relationship between the designated sample objects and the POIs. Additionally, it acquires sample images within a second preset range surrounding the POIs, and extracts sample image location features from these images to characterize the relative positional relationship between the POIs and their surrounding features.

[0161] A prediction model is trained using sample data (including sample geographic location features and sample image location features). After model training, data to be predicted is acquired. This data includes geographic location features representing the relative positional relationship between a specified object within a first preset range around the predicted point of interest and the predicted point of interest itself. It also includes image location features extracted from images covering the predicted point of interest and its surrounding features within a second preset range, representing the relative positional relationship between the predicted point of interest and its surrounding features. The trained prediction model is then used to predict the accuracy of the predicted point of interest's location information based on the data to be predicted.

[0162] Figure 5A A structural block diagram of a data processing apparatus according to an embodiment of the present disclosure is shown. This apparatus can be implemented as part or all of an electronic device through software, hardware, or a combination of both.

[0163] like Figure 5A As shown, the data processing device 510 includes a first acquisition module 511, a second acquisition module 512, and a prediction module 513.

[0164] The first acquisition module 511 is configured to acquire the location information of points of interest;

[0165] The second acquisition module 512 is configured to acquire, based on the location information of the point of interest, the location information of a specified object located within a first preset range around the point of interest and an image of ground features covering the point of interest and its surrounding second preset range.

[0166] The prediction module 513 is configured to predict whether the location information of the point of interest is accurate based on the location information of the point of interest, the location information of the specified object, and the image, using a trained prediction model.

[0167] According to embodiments of this disclosure, the location information of the designated object located within a first preset range around the point of interest includes at least one of the following:

[0168] Location information of buildings within a first preset range around the point of interest;

[0169] Location information of roads within a first preset range around the point of interest;

[0170] Location information of water systems within a first preset range around the point of interest;

[0171] Location information of other points of interest located within a first preset range around the point of interest;

[0172] Location information of objects associated with the point of interest within a first preset range around the point of interest.

[0173] According to embodiments of this disclosure, wherein:

[0174] The location information of the building includes the location information of multiple sampling points around the horizontal outline of the building;

[0175] The location information of the road includes the location information of multiple sampling points along the road;

[0176] The location information of the water system includes the location information of multiple sampling points along the water system;

[0177] The location information of the object associated with the point of interest includes any one of the following: location information of the navigation destination, location information of the device accessing the local network of the point of interest, and location information of the address associated with the point of interest.

[0178] According to embodiments of this disclosure, the step of predicting the accuracy of the location information of the interest point using a trained prediction model based on the location information of the interest point, the location information of the specified object, and the image includes:

[0179] Based on the location information of the specified object and the location information of the point of interest, extract geographic location features to characterize the relative positional relationship between the specified object and the point of interest;

[0180] From the image, extract image location features that characterize the relative positional relationship between the point of interest and its surrounding features;

[0181] The geographic location features and image location features are input into the trained prediction model to predict whether the location information of the point of interest is accurate.

[0182] According to embodiments of this disclosure, the device 510 further includes:

[0183] The third acquisition module 514 is configured to acquire the location information of the sample interest points and the discrimination information on whether the location information of the sample interest points is accurate.

[0184] The fourth acquisition module 515 is configured to acquire, based on the location information of the sample interest point, the location information of the specified sample object located within a first preset range around the sample interest point and the image of the sample covering the sample interest point and its surrounding second preset range.

[0185] The training module 516 is configured to train the prediction model based on the location information of the sample interest points, the location information of the specified object in the sample, the sample image, and the discrimination information on whether the location information of the sample interest points is accurate.

[0186] Figure 5B A structural block diagram of a predictive model training apparatus according to an embodiment of the present disclosure is shown. This apparatus can be implemented as part or all of an electronic device through software, hardware, or a combination of both.

[0187] like Figure 5B As shown, the prediction model training device 520 includes a third acquisition module 514, a fourth acquisition module 515, and a training module 516.

[0188] The third acquisition module 514 is configured to acquire the location information of the sample interest points and the discrimination information on whether the location information of the sample interest points is accurate.

[0189] The fourth acquisition module 515 is configured to acquire, based on the location information of the sample interest point, the location information of the specified sample object within a first preset range around the sample interest point and the image of the sample object within a second preset range covering the sample interest point and its surrounding area.

[0190] The training module 516 is configured to train the prediction model based on the location information of the sample interest points, the location information of the specified object in the sample, the sample image, and the discrimination information on whether the location information of the sample interest points is accurate.

[0191] According to embodiments of this disclosure, the location information of the designated object of the sample located within a first preset range around the sample point of interest includes at least one of the following:

[0192] Location information of buildings within a first preset range surrounding the sample point of interest;

[0193] Location information of roads within a first preset range around the sample point of interest;

[0194] Location information of water systems within a first preset range around the sample point of interest;

[0195] Location information of other points of interest located within a first preset range around the sample point of interest;

[0196] Location information of objects associated with the sample point of interest within a first preset range around the sample point of interest.

[0197] According to embodiments of this disclosure, training the prediction model based on the location information of the sample interest points, the location information of the specified object in the sample, the sample image, and the discrimination information regarding the accuracy of the location information of the sample interest points includes:

[0198] Based on the location information of the specified object in the sample and the location information of the point of interest in the sample, extract sample geographic location features to characterize the relative positional relationship between the specified object in the sample and the point of interest in the sample of this type;

[0199] From the sample images, extract sample image location features to characterize the relative positional relationship between the sample point of interest and its surrounding land features;

[0200] The geographical location features and image location features of the samples are input into the prediction model, and the prediction model is trained with the goal of minimizing the loss function between the output of the prediction model and the discrimination information.

[0201] This disclosure also discloses an electronic device, Figure 6 A structural block diagram of an electronic device according to an embodiment of the present disclosure is shown.

[0202] like Figure 6 As shown, the electronic device 600 includes a memory 601 and a processor 602, wherein the memory 601 is used to store one or more computer instructions, wherein the one or more computer instructions are executed by the processor 602 to implement the method according to embodiments of the present disclosure.

[0203] This disclosure provides a data processing method, including:

[0204] Obtain the location information of points of interest;

[0205] Based on the location information of the point of interest, obtain the location information of a specified object located within a first preset range around the point of interest and an image of ground features covering the point of interest and its surrounding second preset range;

[0206] Based on the location information of the point of interest, the location information of the specified object, and the image, a trained prediction model is used to predict whether the location information of the point of interest is accurate.

[0207] According to embodiments of this disclosure, the location information of the designated object located within a first preset range around the point of interest includes at least one of the following:

[0208] Location information of buildings within a first preset range around the point of interest;

[0209] Location information of roads within a first preset range around the point of interest;

[0210] Location information of water systems within a first preset range around the point of interest;

[0211] Location information of other points of interest located within a first preset range around the point of interest;

[0212] Location information of objects associated with the point of interest within a first preset range around the point of interest.

[0213] According to embodiments of this disclosure, wherein:

[0214] The location information of the building includes the location information of multiple sampling points around the horizontal outline of the building;

[0215] The location information of the road includes the location information of multiple sampling points along the road;

[0216] The location information of the water system includes the location information of multiple sampling points along the water system;

[0217] The location information of the object associated with the point of interest includes any one of the following: location information of the navigation destination, location information of the device accessing the local network of the point of interest, and location information of the address associated with the point of interest.

[0218] According to embodiments of this disclosure, the step of predicting the accuracy of the location information of the interest point using a trained prediction model based on the location information of the interest point, the location information of the specified object, and the image includes:

[0219] Based on the location information of the specified object and the location information of the point of interest, extract geographic location features to characterize the relative positional relationship between the specified object and the point of interest;

[0220] From the image, extract image location features that characterize the relative positional relationship between the point of interest and its surrounding features;

[0221] The geographic location features and image location features are input into the trained prediction model to predict whether the location information of the point of interest is accurate.

[0222] According to embodiments of this disclosure, the method further includes:

[0223] Acquire the location information of the sample interest points and the information on whether the location information of the sample interest points is accurate;

[0224] Based on the location information of the sample point of interest, obtain the location information of the specified sample object within a first preset range around the sample point of interest and the sample image with a shooting range covering the sample point of interest and its surrounding second preset range;

[0225] The prediction model is trained based on the location information of the sample interest points, the location information of the specified object in the sample, the sample image, and the discrimination information on whether the location information of the sample interest points is accurate.

[0226] This disclosure also provides a method for training a prediction model, including:

[0227] Acquire the location information of the sample interest points and the information on whether the location information of the sample interest points is accurate;

[0228] Based on the location information of the sample point of interest, obtain the location information of the specified sample object within a first preset range around the sample point of interest and the sample image with a shooting range covering the sample point of interest and its surrounding second preset range;

[0229] The prediction model is trained based on the location information of the sample interest points, the location information of the specified object in the sample, the sample image, and the discrimination information on whether the location information of the sample interest points is accurate.

[0230] According to embodiments of this disclosure, the location information of the designated object of the sample located within a first preset range around the sample point of interest includes at least one of the following:

[0231] Location information of buildings within a first preset range surrounding the sample point of interest;

[0232] Location information of roads within a first preset range around the sample point of interest;

[0233] Location information of water systems within a first preset range around the sample point of interest;

[0234] Location information of other points of interest located within a first preset range around the sample point of interest;

[0235] Location information of objects associated with the sample point of interest within a first preset range around the sample point of interest.

[0236] According to embodiments of this disclosure, training the prediction model based on the location information of the sample interest points, the location information of the specified object in the sample, the sample image, and the discrimination information regarding the accuracy of the location information of the sample interest points includes:

[0237] Based on the location information of the specified object in the sample and the location information of the point of interest in the sample, extract sample geographic location features to characterize the relative positional relationship between the specified object in the sample and the point of interest in the sample of this type;

[0238] From the sample images, extract sample image location features to characterize the relative positional relationship between the sample point of interest and its surrounding land features;

[0239] The geographical location features and image location features of the samples are input into the prediction model, and the prediction model is trained with the goal of minimizing the loss function between the output of the prediction model and the discrimination information.

[0240] Figure 7 A schematic diagram of the structure of a computer system suitable for implementing the method according to embodiments of the present disclosure is shown.

[0241] like Figure 7 As shown, the computer system 700 includes a processing unit 701, which can execute various processes described in the above embodiments according to a program stored in a read-only memory (ROM) 702 or a program loaded from a storage section 708 into a random access memory (RAM) 703. The RAM 703 also stores various programs and data required for the operation of the system 700. The processing unit 701, ROM 702, and RAM 703 are interconnected via a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0242] The following components are connected to the I / O interface 705: an input section 706 including a keyboard, mouse, etc.; an output section 707 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 708 including a hard disk, etc.; and a communication section 709 including a network interface card such as a LAN card, modem, etc. The communication section 709 performs communication processing via a network such as the Internet. A drive 710 is also connected to the I / O interface 705 as needed. A removable medium 711, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on the drive 710 as needed so that computer programs read from it can be installed into the storage section 708 as needed. The processing unit 701 can be implemented as a CPU, GPU, TPU, FPGA, NPU, etc.

[0243] In particular, according to embodiments of this disclosure, the methods described above can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program tangibly embodied on a computer-readable storage medium, the computer program containing program code for performing the methods described above. In such embodiments, the computer program can be downloaded and installed from a network via communication section 709, and / or installed from removable medium 711.

[0244] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0245] The units or modules described in the embodiments of this disclosure can be implemented in software or programmable hardware. The described units or modules can also be located in a processor, and the names of these units or modules do not necessarily constitute a limitation on the unit or module itself.

[0246] In another aspect, this disclosure also provides a computer-readable storage medium, which may be a computer-readable storage medium included in the electronic device or computer system described above; or it may be a standalone computer-readable storage medium not assembled into a device. The computer-readable storage medium stores one or more programs, which are used by one or more processors to perform the methods described in this disclosure.

[0247] The above description is merely a preferred embodiment of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features disclosed in this disclosure that have similar functions.

Claims

1. A data processing method, comprising: Obtain the location information of points of interest; Based on the location information of the point of interest, obtain the location information of a specified object located within a first preset range around the point of interest and an image of ground features covering the point of interest and its surrounding second preset range; Based on the location information of the point of interest, the location information of the specified object, and the image, a trained prediction model is used to predict whether the location information of the point of interest is accurate. The training of the prediction model includes: Based on the location information of the specified object and the location information of the point of interest, extract sample geographic location features to characterize the relative positional relationship between the specified object and the point of interest. Extract sample image location features from the sample images to characterize the relative positional relationship between the sample point of interest and its surrounding land features; The geographical location features of the samples and the location features of the sample images are input into the prediction model. The prediction model is trained with the goal of minimizing the loss function between the output of the prediction model and the discrimination information on whether the location information of the sample interest points is accurate.

2. The method according to claim 1, wherein, The location information of the designated object located within a first preset range around the point of interest includes at least one of the following: Location information of buildings within a first preset range around the point of interest; Location information of roads within a first preset range around the point of interest; Location information of water systems within a first preset range around the point of interest; Location information of other points of interest located within a first preset range around the point of interest; Location information of objects associated with the point of interest within a first preset range around the point of interest.

3. The method according to claim 2, wherein: The location information of the building includes the location information of multiple sampling points around the horizontal outline of the building; The location information of the road includes the location information of multiple sampling points along the road; The location information of the water system includes the location information of multiple sampling points along the water system; The location information of the object associated with the point of interest includes any one of the following: location information of the navigation destination, location information of the device accessing the local network of the point of interest, and location information of the address associated with the point of interest.

4. The method according to claim 1, wherein, The step of using a trained prediction model to predict the accuracy of the location information of the interest point based on the location information of the specified object and the image includes: Based on the location information of the specified object and the location information of the point of interest, extract geographic location features to characterize the relative positional relationship between the specified object and the point of interest; From the image, extract image location features that characterize the relative positional relationship between the point of interest and its surrounding features; The geographic location features and image location features are input into the trained prediction model to predict whether the location information of the point of interest is accurate.

5. The method according to claim 1, further comprising: Acquire the location information of the sample interest points and the information on whether the location information of the sample interest points is accurate; Based on the location information of the sample point of interest, obtain the location information of the specified sample object within a first preset range around the sample point of interest and the sample image with a shooting range covering the sample point of interest and its surrounding second preset range; The prediction model is trained based on the location information of the sample interest points, the location information of the specified object in the sample, the sample image, and the discrimination information on whether the location information of the sample interest points is accurate.

6. A method for training a prediction model, comprising: Acquire the location information of the sample interest points and the information on whether the location information of the sample interest points is accurate; Based on the location information of the sample point of interest, obtain the location information of the specified sample object within a first preset range around the sample point of interest and the sample image with a shooting range covering the sample point of interest and its surrounding second preset range; Based on the location information of the sample points of interest, the location information of the sample designated object, the sample image, and the discrimination information regarding the accuracy of the location information of the sample points of interest, the prediction model is trained, including: extracting sample geographic location features to characterize the relative positional relationship between the sample designated object and the sample points of interest based on the location information of the sample designated object and the sample points of interest; extracting sample image location features from the sample image to characterize the relative positional relationship between the sample points of interest and their surrounding features; inputting the sample geographic location features and sample image location features into the prediction model, and training the prediction model with the objective of minimizing the loss function between the output of the prediction model and the discrimination information.

7. The method according to claim 6, wherein, The location information of the specified object of the sample located within a first preset range around the sample point of interest includes at least one of the following: Location information of buildings within a first preset range surrounding the sample point of interest; Location information of roads within a first preset range around the sample point of interest; Location information of water systems within a first preset range around the sample point of interest; Location information of other points of interest located within a first preset range around the sample point of interest; Location information of objects associated with the sample point of interest within a first preset range around the sample point of interest.

8. An electronic device comprising a memory and a processor; wherein, The memory is used to store one or more computer instructions, wherein the one or more computer instructions are executed by the processor to implement the steps of the method according to any one of claims 1-7.

9. A computer program product comprising computer instructions that, when executed by a processor, implement the steps of the method according to any one of claims 1-7.

Images