Method and apparatus for determining a region of interest

Abstract

Embodiments of the present specification provide a method and device for determining an area of interest. In the method for determining an area of interest, a remote sensing image containing a target point of interest is acquired; the remote sensing image containing the target point of interest is provided to an area of interest determination model to obtain a target number of area of interest boundary prediction points; a to-be-determined area of an area of interest image containing the target point of interest is determined according to the target number of area of interest boundary prediction points; and an area of interest containing the target point of interest is determined according to the to-be-determined area of the area of interest image. Thus, an area of interest containing a point of interest can be more accurately determined.

Description

Technical Field

[0001] The embodiments in this specification generally relate to the field of computer technology, and more particularly to methods and apparatus for determining regions of interest. Background Technology

[0002] With the rapid development of artificial intelligence (AI) technology, its application in assisting geographic information collection has become increasingly widespread. Point of Interest (POI) is a concept within Geographic Information Systems (GIS) that refers to geographic objects that can be abstracted as points. These typically include geographic entities closely related to people's lives, such as schools, banks, restaurants, gas stations, hospitals, supermarkets, shopping malls, and bus stops. In some cases, we often need to know the specific extent of an Area of ​​Interest (AOI) that contains POIs.

[0003] One existing technique often uses pixel recognition of remote sensing images (RSI) to identify geographic features (such as a school playground) within a region of interest (ROI), thereby delineating the approximate extent of the ROI. However, this approach struggles to identify insignificant geographic features (such as buildings and green areas in residential areas and schools), thus failing to accurately determine the boundaries of the ROI. Another existing technique uses road networks or vehicle trajectories to determine the internal grid of a tic-tac-toe road network, identifying the grid containing points of interest as the ROI. However, detailed road network data (including internal roads and highways) is difficult to obtain, and the true outline of the ROI is not necessarily bounded by roads, making accurate identification of the ROI's outline challenging. Therefore, a method for determining ROI that can at least partially overcome these shortcomings is needed. Summary of the Invention

[0004] In view of the above, embodiments of this specification provide a method and apparatus for determining regions of interest. Using this method and apparatus, regions of interest containing points of interest can be determined more accurately.

[0005] According to one aspect of an embodiment of this specification, a method for determining a region of interest is provided, comprising: acquiring a remote sensing image containing a target point of interest; providing the remote sensing image containing the target point of interest to a region of interest determination model to obtain a target number of predicted boundary points of the region of interest; determining a region of interest image to be determined containing the target point of interest based on the target number of predicted boundary points of the region of interest; and determining the region of interest containing the target point of interest based on the region of interest image to be determined.

[0006] According to another aspect of the embodiments of this specification, a model training method for dialogue-based item recommendation is provided, comprising: an image acquisition unit configured to acquire a remote sensing image containing a target point of interest; a boundary point prediction unit configured to provide the remote sensing image containing the target point of interest to a region of interest determination model to obtain a target number of boundary prediction points for regions of interest; an image region to be determined unit configured to determine a region of interest image to be determined containing the target point of interest based on the target number of boundary prediction points for regions of interest; and a region of interest determination unit configured to determine a region of interest containing the target point of interest based on the region of interest image to be determined.

[0007] According to another aspect of the embodiments of this specification, an apparatus for determining a region of interest is provided, comprising: at least one processor, a memory coupled to the at least one processor, and a computer program stored on the memory, wherein the at least one processor executes the computer program to implement the method for determining a region of interest as described above.

[0008] According to another aspect of the embodiments of this specification, a computer-readable storage medium is provided that stores a computer program, which, when executed by a processor, implements the method for determining a region of interest as described above.

[0009] According to another aspect of the embodiments of this specification, a computer program product is provided, including a computer program that is executed by a processor to implement the method for determining a region of interest as described above. Attached Figure Description

[0010] A further understanding of the nature and advantages of this specification can be achieved by referring to the following figures. In the figures, similar components or features may have the same reference numerals.

[0011] Figure 1 An exemplary architecture of a method and apparatus for determining a region of interest according to embodiments of this specification is shown.

[0012] Figure 2 A flowchart illustrating an example of a method for determining a region of interest according to an embodiment of this specification is shown.

[0013] Figure 3 A schematic diagram illustrating an example of the process for obtaining the target number of region of interest boundary prediction points according to an embodiment of this specification.

[0014] Figure 4 A schematic diagram illustrating an example of a process for obtaining training samples according to an embodiment of this specification is shown.

[0015] Figure 5 A flowchart illustrating an example of a process for determining a region of interest containing a target point of interest according to an embodiment of this specification is shown.

[0016] Figure 6 A schematic diagram illustrates yet another example of the process for determining a region of interest containing a target point of interest according to an embodiment of this specification.

[0017] Figure 7 A flowchart illustrating an example of the process for determining the reserved boundary of a region of interest road network to be determined according to an embodiment of this specification is shown.

[0018] Figure 8 A schematic diagram illustrates yet another example of the process for determining an area of ​​an urban functional zone containing a target point of interest, according to an embodiment of this specification.

[0019] Figure 9 A flowchart illustrating yet another example of a method for determining a region of interest according to an embodiment of this specification is shown.

[0020] Figure 10 A block diagram illustrating an example of a device for determining a region of interest according to an embodiment of this specification is shown.

[0021] Figure 11 A block diagram of an example of an interest region determination unit in an apparatus for determining an interest region according to an embodiment of this specification is shown.

[0022] Figure 12 A block diagram of an example of a region fusion module in an apparatus for determining a region of interest according to an embodiment of this specification is shown.

[0023] Figure 13 A block diagram illustrating yet another example of a device for determining a region of interest according to an embodiment of this specification is shown.

[0024] Figure 14 A block diagram illustrating an example of a device for determining a region of interest according to an embodiment of this specification is shown. Detailed Implementation

[0025] The subject matter described herein will be discussed below with reference to exemplary embodiments. It should be understood that these embodiments are discussed merely to enable those skilled in the art to better understand and implement the subject matter described herein, and are not intended to limit the scope, applicability, or examples set forth in the claims. The function and arrangement of the elements discussed may be changed without departing from the scope of the embodiments described herein. Various processes or components may be omitted, substituted, or added as needed in the various examples. Furthermore, features described in some examples may be combined in other examples.

[0026] As used herein, the term "comprising" and its variations are open terms meaning "including but not limited to". The term "based on" means "at least partially based on". The terms "one embodiment" and "an embodiment" mean "at least one embodiment". The term "another embodiment" means "at least one other embodiment". The terms "first", "second", etc., may refer to different or the same objects. Other definitions, whether explicit or implicit, may be included below. Unless explicitly indicated by the context, the definition of a term shall remain consistent throughout the specification.

[0027] The method and apparatus for determining a region of interest according to embodiments of this specification will now be described in detail with reference to the accompanying drawings.

[0028] Figure 1 An exemplary architecture 100 of a method and apparatus for determining a region of interest according to embodiments of this specification is shown.

[0029] exist Figure 1 In this context, network 110 is used to interconnect terminal device 120 and application server 130.

[0030] Network 110 can be any type of network capable of interconnecting network entities. Network 110 can be a single network or a combination of various networks. In terms of coverage, network 110 can be a local area network (LAN), a wide area network (WAN), etc. In terms of the carrying medium, network 110 can be a wired network, a wireless network, etc. In terms of data switching technology, network 110 can be a circuit-switched network, a packet-switched network, etc.

[0031] Terminal device 120 can be any type of electronic computing device capable of connecting to network 110, accessing servers or websites on network 110, processing data or signals, etc. For example, terminal device 120 can be a desktop computer, laptop computer, tablet computer, smartphone, drone, etc. Although in Figure 1 Only one terminal device is shown in the diagram, but it should be understood that a different number of terminal devices may be connected to network 110.

[0032] In one implementation, terminal device 120 can be used by a user. Terminal device 120 may include an application client (e.g., application client 121) that can provide various services to the user. In one example, the application client may be a shopping application, a photo application, a map application, etc. In some cases, application client 121 may interact with application server 130. For example, application client 121 may transmit remote sensing images containing points of interest to application server 130. However, it should be understood that in other cases, application server 130 may also obtain remote sensing images containing points of interest locally, without interacting with application client 121.

[0033] Application server 130 can connect to interest area database 140. Interest area database 140 can include information on various interest areas containing different points of interest. In one example, interest area database 140 can include the boundary coordinates of urban functional areas (e.g., schools, hospitals, etc.). In another example, interest area database 140 can include the locations of scenic spots (e.g., parks, attractions).

[0034] It should be understood that Figure 1 All network entities shown are exemplary, and any other network entities may be involved in Architecture 100 depending on the specific application requirements.

[0035] Figure 2 A flowchart of a method 200 for determining a region of interest according to an embodiment of this specification is shown.

[0036] like Figure 2 As shown, at 210, a remote sensing image containing the target point of interest is acquired.

[0037] In this embodiment, the aforementioned remote sensing images can refer to films or photographs recording the electromagnetic wave magnitudes of various ground features, and may include aerial photographs (e.g., photographs taken from various aircraft or using drones) and satellite photographs. In one example, the aforementioned target points of interest can be pre-defined points of interest, such as XX School, YY Park, etc. In another example, the aforementioned target points of interest can also be a type of point of interest that meets preset requirements. The preset requirements may, for example, be having an independent geographical location. Thus, places such as schools, hospitals, residential areas, industrial parks, and science parks can be considered as the aforementioned target points of interest, while ground-floor shops such as barbershops and milk tea shops, or department store counters without independent stores, although they can also be considered points of interest, do not belong to the aforementioned target points of interest.

[0038] In this embodiment, optionally, the coverage area of ​​the remote sensing image can also be determined based on the road network grid where the target point of interest is located. For example, the remote sensing image may be an image taken for the road network grid where the target point of interest is located.

[0039] In this embodiment, the remote sensing image can be associated with the location information of the included target points of interest. In one example, the location information of the target points of interest can be obtained in advance through location-based services (LBS). In another example, the location information of the target points of interest can be retrieved from a pre-defined geographic information database.

[0040] In step 220, the remote sensing image containing the target interest points is provided to the interest region determination model to obtain the target number of interest region boundary prediction points.

[0041] In this embodiment, a remote sensing image containing target points of interest can be provided to a region of interest (ROI) determination model to obtain a target number of ROI boundary prediction points. The ROI determination model can be any machine learning model used to identify ROI boundary points. The ROI determination model can be trained using supervised learning with training samples. The training samples can include sample remote sensing images and sample annotation information. The sample remote sensing images can display sample points of interest (e.g., schools, parks, hospitals, etc.). The sample annotation information can include multiple labeled boundary points located on the boundaries of the areas where the sample points of interest are located (e.g., school walls, hospital walls, the boundary between a park and an external road).

[0042] In one example, the region of interest (ROI) determination model described above could be based on a Convolutional Neural Network (CNN). In another example, it could be based on a Transformer. For instance, the ROI determination model could include a feature extraction network, a Transformer encoder, a Transformer decoder, and a feedforward neural network (FNN). The feature extraction network could be various backbone networks used for extracting image features, such as ResNet, VGG, etc.

[0043] In one example, reference Figure 3 , Figure 3 An example of the process for obtaining the target number of predicted boundary points of a region of interest according to an embodiment of this specification is shown. Figure 3As shown, a remote sensing image 310 containing the target interest point 311 can be provided to the feature extraction network 320 to obtain initial image features corresponding to the remote sensing image 310. Then, a first image feature can be obtained by performing token projection 330 on the initial image features. The token projection 330 may include, for example, convolutional processing. The obtained first image feature is fused with a positional encoding to obtain image encoding feature 331. The positional encoding is used to represent the position of pixels in the first image feature. The image encoding feature 331 is provided to the Transformer encoder 350 to obtain a corresponding encoding 351. A second image feature can also be obtained by performing query projection 340 on the initial image features. The query projection 340 may include, for example, convolutional processing (usually using different parameters than the token projection 330, i.e., convolutional kernel). Then, based on the position of the target interest point 311, the feature vector 341 corresponding to the target interest point 311 is extracted from the second image features. The image encoding features 331 obtained by the Transformer encoder 350 and the feature vector 341 corresponding to the target interest point 311 are jointly provided to the Transformer decoder 360 to obtain the decoded features 361. Finally, the decoded features 361 are provided to the feedforward neural network (FFN) to obtain multiple predicted boundary points (as shown by the white dots in image 370).

[0044] Optionally, refer to the following: Figure 4 , Figure 4 A schematic diagram illustrating an example of a process for obtaining training samples according to an embodiment of this specification is shown.

[0045] like Figure 4 As shown, in this embodiment, sample remote sensing images (such as...) can be acquired first. Figure 4 As shown in Figure 410). The aforementioned remote sensing image can display the sample boundary region containing the sample interest points and the location of the sample interest points. The aforementioned sample boundary region is usually the actual area covered by the aforementioned interest region containing the sample interest points (e.g., Figure 4 As shown in Figure 411). The location of the above sample interest points can be represented by the coordinates of the location of the interest region containing the sample interest points after it has been abstracted into a single point (e.g., Figure 4 (As shown in Figure 412). In one example, the location of the sample interest point can be at the center of the interest region containing the sample interest point.

[0046] In this embodiment, the location of the sample interest point (e.g., Figure 4As shown in Figure 412, the origin of the polar coordinate system is used, and a ray is drawn at preset angles (e.g., ...). Figure 4 As shown in Figure 421 (only a portion is shown in this figure), the preset angle can be, for example, 2π / n, where n is a positive integer. The value of n can be, for example, between 20 and 50. Then, the intersection points of the ray and the sample boundary region can be determined as multiple boundary markers corresponding to the sample remote sensing image (e.g., ...). Figure 4 The white dots within the dashed area shown in Figure 422 (only a portion is shown in this figure) can be used to determine the sample remote sensing image 410 and its corresponding multiple boundary markers as training samples for training the aforementioned region of interest determination model.

[0047] Based on this, this solution provides a training sample acquisition method not recorded in existing technologies, which can effectively select boundary annotation points in sample images, thereby providing a basis for training a model to determine the region of interest.

[0048] In step 230, based on the target number of interest region boundary prediction points, the undetermined region of the interest region image containing the target interest points is determined.

[0049] In this embodiment, based on the target number of region of interest boundary prediction points, the undetermined region of interest containing the target region of interest can be determined in various ways. In one example, the target number of region of interest boundary prediction points can be sequentially connected to form a closed region, which is then determined as the undetermined region of interest containing the target region of interest.

[0050] In step 240, the region of interest containing the target point of interest is determined based on the region of interest to be determined in the region of interest image.

[0051] In this embodiment, based on the region of interest (ROI) image to be determined, the ROI containing the target ROI can be determined in various ways. In one example, the ROI image to be determined can be directly identified as the ROI containing the target ROI. In another example, the ROI image to be determined can be post-processed, and the processed region can be identified as the ROI containing the target ROI. For example, the post-processing may include straightening or smoothing the jagged boundaries.

[0052] Optionally, refer to Figure 5 , Figure 5 A flowchart illustrating an example of a process 500 for determining a region of interest containing a target point of interest according to an embodiment of this specification is shown.

[0053] like Figure 5 As shown in Figure 510, the undetermined area of ​​the road network containing the target point of interest is obtained.

[0054] In this embodiment, the aforementioned region of interest (ROI) road network to be determined can be based on road network information involving the aforementioned target ROI. In one example, the ROI road network to be determined (e.g., a road network grid containing the aforementioned target ROI) can be determined from the road network information based on the target ROI. The aforementioned road network information may include road location information established based on the road network and related information about the grid enclosed by the roads.

[0055] In one example, the aforementioned road network information can be obtained as follows: First, the road network can be filtered by city. Then, the filtered road network is filtered to obtain a filtered road network. This filtering process may include, for example, selecting roads based on type (e.g., highways, elevated roads, national highways, provincial highways, county roads, township roads, main streets, etc.) and removing internal roads and dead-end roads. Next, morphological transformations (e.g., erosion, dilation, and refinement) can be performed on the filtered road network. Then, the morphologically transformed road network is segmented into regions to form individual road network grids. Finally, the resulting road network can be used to process the filtered road network, forming road network information including multiple closed-loop connected regions.

[0056] In step 520, the regions of interest containing the target interest points are determined by fusing the undetermined regions of the region of interest image and the undetermined regions of the road network in the region of interest.

[0057] In this embodiment, the regions of interest (ROI) to be determined, including the image region to be determined and the road network region to be determined, can be fused in various ways to determine the fused region as the ROI containing the target ROI. In one example, the intersection of the image region to be determined and the road network region to be determined can be determined as the ROI containing the target ROI.

[0058] Based on this, this scheme can comprehensively utilize the regions of interest (ROIs) determined by remote sensing images and the regions of interest (ROIs) determined by road networks as the criteria for determining ROIs containing target ROIs, which can effectively improve the accuracy of the determined ROIs containing target ROIs.

[0059] Optionally, further reference Figure 6 , Figure 6 A schematic diagram illustrates yet another example of the process for determining a region of interest containing a target point of interest according to an embodiment of this specification.

[0060] like Figure 6 As shown, in this embodiment, the distance from the target point of interest to at least one target location in the undetermined area of ​​the road network of the interest region can be determined based on the location of the target point of interest. In one example, the target point of interest (e.g., Figure 6 Distances from each target location (as shown in Figure 610) to each target location (e.g.) Figure 6 (As shown in d1 to d5).

[0061] In this embodiment, the reserved boundary of the undetermined area of ​​the region of interest road network (e.g., the area enclosed by boundaries 620, 630, 640, and 650 in the figure) can be determined based on at least one determined distance. In one example, an allowable distance can be determined from the determined distances. This allowable distance can serve as the primary basis for retaining the boundary. As an example, the allowable distance can be the N smallest distances among the determined distances (N is a positive integer). For example, the allowable distances can be d1 to d3. As another example, the allowable distance can be a distance less than a predetermined distance threshold. Then, the boundaries in the undetermined area of ​​the region of interest road network corresponding to the allowable distances can be retained. For example, the reserved boundary corresponding to allowable distance d1 can be boundary 620. The reserved boundary corresponding to allowable distance d3 can be boundary 630. The reserved boundaries corresponding to allowable distance d2 can be boundaries 620 and 630.

[0062] In this embodiment, the reserved boundary determined above can be the complete boundary or a portion of the boundary of the road network to be determined in the region of interest. In one example, the length of the reserved boundary can be determined in various ways. For example, it can utilize a hyperparameter trained to characterize the reserved length or reserved ratio, or it can be matched with the aforementioned allowable distance. For example, the reserved length of reserved boundary 620 can be twice the allowable distance d3. As another example, the reserved length of reserved boundary 630 can be twice the allowable distance d1. Yet another example, the reserved length of reserved boundary 630 can be the sum of the allowable distance d1 and the aforementioned predetermined distance threshold.

[0063] In this embodiment, based on the region of interest image to be determined (e.g., Figure 6 (as shown in 660) and reserved boundaries (e.g.) Figure 6 As shown in Figures 620 and 630, the region of interest containing the target point of interest can be determined in various ways. In one example, a rectangular region 670 can be constructed based on the length of the reserved boundary. In this example, when the region of interest to be determined 660 does not exceed the rectangular region 670, and the boundary of the region of interest to be determined 660 intersects with the boundary of the rectangular region 670, the rectangular region 670 can be determined as the region of interest containing the target point of interest.

[0064] Based on this, this solution provides a method for determining the retention boundary based on the undetermined area of ​​the road network in the region of interest, and then determining the region of interest containing the target point of interest based on the undetermined area of ​​the image of the region of interest and the retention boundary. This provides a specific fusion scheme for the undetermined area of ​​the image of the region of interest and the undetermined area of ​​the road network in the region of interest, which helps to improve the accuracy of determining the region of interest containing the target point of interest.

[0065] Optionally, refer to the following: Figure 7 , Figure 7 A flowchart illustrating an example of a process 700 for determining the reserved boundary of a region of interest road network to be determined according to an embodiment of this specification is shown.

[0066] like Figure 7 As shown in Figure 710, the minimum circumscribed circle radius of the region of interest to be determined in the image is determined.

[0067] In this embodiment, the minimum circumscribed circle radius of the region of interest to be determined in the image can be determined by the number of target region boundary prediction points obtained from the region of interest determination model.

[0068] In 720, the boundary of the undetermined area of ​​the road network involving at least one determined distance that is not greater than the minimum circumscribed circle radius is determined as the boundary to be retained.

[0069] In this embodiment, at least one determined distance can be compared with the aforementioned minimum circumscribed circle radius. Then, the boundary of the undetermined area of ​​the road network corresponding to a distance not greater than the aforementioned minimum circumscribed circle radius can be determined as the boundary to be retained. In one example, the correspondence between the aforementioned distance and boundary can be referred to the foregoing. Figure 6 The corresponding descriptions in the embodiments will not be repeated here.

[0070] At 730, extract the reserved boundaries of the undetermined areas of the road network in the region of interest from the boundaries to be reserved.

[0071] In this embodiment, reference can be made to the foregoing. Figure 6 The corresponding description in the embodiments involves extracting all or part of the aforementioned boundaries to be retained, thereby obtaining the retained boundaries of the undetermined area of ​​the road network in the region of interest.

[0072] Based on this, this scheme can combine the minimum circumscribed circle radius of the undetermined region of the region of interest image determined by the target number of region of interest boundary prediction points obtained from the region of interest determination model, and determine the reserved boundary in the undetermined region of the region of interest network determined by the road network. This effectively utilizes the most reliable parts of the remote sensing image method and the road network method and then fuses them, thereby helping to improve the accuracy of the determined region of interest.

[0073] See also Figure 8 , Figure 8 A schematic diagram illustrates yet another example of the process for determining a region of interest containing a target point of interest according to an embodiment of this specification.

[0074] like Figure 8As shown, in one example, the undetermined region of interest (ROI) road network containing point of interest 810 can be the region enclosed by boundaries 820, 830, 840, and 850. Boundaries 820 and 830 can be preserved boundaries of the undetermined ROI road network. Boundaries 840 and 850 can be non-preserved boundaries of the undetermined ROI road network. The undetermined region of interest (ROI) image can be determined (e.g.,...). Figure 8 The circumscribed polygon (e.g., shown in Figure 860) is a bounded polygon. Figure 8 (As shown in 870).

[0075] In one example, a rectangular region 880 can be constructed symmetrically based on the length of the aforementioned reserved boundary. In this example, when the boundary 881 of the rectangular region 880 intersects with the region of interest to be determined 860 of the image, it indicates that the constructed rectangular region 880 does not completely encompass the region of interest to be determined. Therefore, the constructed rectangular region 880 can be expanded based on the aforementioned circumscribed polygon 870, and the expanded region (such as...) can be... Figure 8 The union of regions 880 and 890 is determined as the region of interest containing the target point of interest.

[0076] Based on this, this solution can determine the circumscribed polygon of the region of interest image based on the non-reserved boundary determined by the region of interest road network to be determined region, and then determine the region of interest containing the target point of interest based on the boundary of the circumscribed polygon and the reserved boundary, thereby enriching the fusion method of region of interest image to be determined region and region of interest road network to be determined region, and helping to improve the accuracy of determining the region of interest containing the target point of interest in specific scenarios.

[0077] Continue to refer to Figure 9 , Figure 9 A flowchart is shown as yet another example of a method 900 for determining a region of interest according to an embodiment of this specification.

[0078] like Figure 9 As shown, at 910, an image containing the target interest point is acquired.

[0079] In step 920, the image containing the target interest points is provided to the interest region determination model to obtain the target number of interest region boundary prediction points.

[0080] In this embodiment, the aforementioned region of interest may include urban functional zones. In this embodiment, an urban functional zone can refer to the spatial distribution of various functional activities within a city and the corresponding sub-district differentiation. In simpler terms, an urban functional zone can be the smallest geographical functional block unit within a city, representing a geographical entity. In one example, an urban functional zone can be a residential area, school, hospital, park, industrial park, science park, etc.

[0081] In step 930, based on the target number of interest region boundary prediction points, the undetermined region of the interest region image containing the target interest points is determined.

[0082] At 940, the region of interest containing the target point of interest is determined based on the region of interest to be determined in the region of interest image.

[0083] It should be noted that steps 910-940 above can be referred to the aforementioned... Figure 2 The relevant descriptions of steps 210-240 in the embodiments will not be repeated here.

[0084] At 950, obtain the feature vector corresponding to the determined urban functional area.

[0085] In this embodiment, feature information corresponding to the determined urban functional area can be obtained through various methods. This feature information can include various information that can be used to assess the accuracy of the determined urban functional area. In one example, the feature information may include, but is not limited to, at least one of the following: geometric topology matching information, population activity distribution patterns, address keyword matching information, WIFI connection number distribution patterns, and payment transaction behavior data. Specifically, the geometric topology matching information can be used to characterize the matching degree between the undetermined area of ​​interest (ROI) image and the undetermined area of ​​ROI road network corresponding to the determined urban functional area (e.g., the ratio of the intersection to the union of the undetermined ROI image and the undetermined ROI road network). The population activity distribution patterns can be used to characterize population density. For example, whether the peak population density is on weekdays or weekends, during working hours (e.g., 9:00-17:00) or non-working hours (e.g., 19:00-3:00). The address keyword matching information can be used to characterize the matching degree between the associated address and the determined urban functional area. The associated address may be, for example, the receiving or sending address in logistics information or other POI information near the determined urban functional area. The distribution patterns of the number of WIFI connections and payment transaction data mentioned above can be similar to the distribution patterns of population activities mentioned above, and can be used to characterize the population density at different times.

[0086] In this embodiment, the corresponding feature vector can be determined based on the feature information obtained above that corresponds to the determined urban functional area. In one example, the above feature information can be converted into corresponding codes. For example, the feature vector determined by the feature information corresponding to the determined urban functional area can be [0.7,1,1,0.85,1,1]. Each of the above elements can be used to characterize the matching degree between the undetermined area of ​​the region of interest image and the undetermined area of ​​the road network corresponding to the determined urban functional area, the peak of population density on weekdays, the peak of population density during working hours, the matching degree between the associated address and the determined urban functional area, the peak of the number of WIFI connections during working hours, and the peak of payment transaction data during working hours.

[0087] In step 960, the feature vector is provided to the urban functional area assessment model to obtain the accuracy of the corresponding urban functional area determination.

[0088] In this embodiment, feature vectors can be provided to the urban functional area assessment model to obtain the accuracy of the corresponding urban functional area determination. The urban functional area assessment model can be trained using machine learning methods with training samples containing sample urban functional area regions and accuracy labels (e.g., 0 or 1). In one example, the urban functional area assessment model can be a decision tree-based model (e.g., random forest, XGBoost, etc.).

[0089] At 970, in response to the determination that the accuracy is greater than a predetermined threshold, the urban functional area information is stored in a preset database.

[0090] In this embodiment, when the accuracy obtained in step 960 is greater than a predetermined threshold (e.g., 0.7), the urban functional area information corresponding to the accuracy can be stored in a preset database. The preset database can be any database used to store information related to urban functional areas.

[0091] Based on this, this solution can use the urban functional area assessment model to assess the accuracy of the urban functional areas determined by remote sensing images or road network fusion, and store the urban functional area determination information that meets the accuracy requirements in the database, providing data and technical foundation for various downstream urban services such as food delivery / express delivery, commercial site selection, and location-based service precision marketing.

[0092] use Figures 1-9The method disclosed herein for determining regions of interest (ROIs) utilizes an ROI determination model to obtain a target number of ROI boundary prediction points in a remote sensing image containing target ROIs. Furthermore, based on the obtained target number of ROI boundary prediction points, the method further determines the ROI containing the target ROIs within the undetermined region of the ROI image. Compared to existing methods of identifying geographic features in ROIs using remote sensing image pixel recognition, this scheme identifies features from boundary prediction points rather than from regional geographic features or roads. This makes it applicable to boundary identification of various types of ROIs (e.g., ROIs without landmark elements such as playgrounds, ROIs not demarcated by roads, etc.), and it does not rely on the tic-tac-toe grid formed by road network data, improving the applicability and accuracy of the scheme.

[0093] Figure 10 A block diagram illustrating an example of a device 1000 for determining a region of interest according to an embodiment of this specification is shown. This device embodiment can be used with... Figures 1-9 Corresponding to the method embodiments shown, this device can be specifically applied to various electronic devices.

[0094] like Figure 10 As shown, the apparatus 1000 for determining the region of interest may include an image acquisition unit 1010, a boundary point prediction unit 1020, an image region to be determined determination unit 1030, and a region of interest determination unit 1040.

[0095] Image acquisition unit 1010 is configured to acquire remote sensing images containing target points of interest. The operation of image acquisition unit 1010 can be referred to above. Figure 2 The operation described in section 210.

[0096] Boundary point prediction unit 1020 is configured to provide the remote sensing image containing the target interest points to the region of interest determination model to obtain a target number of interest region boundary prediction points. The operation of boundary point prediction unit 1020 can be referred to above. Figure 2 The operation described in section 220.

[0097] In one example, the specific structure of the above-mentioned region of interest determination model can be referred to the foregoing. Figure 3 In the embodiments Figure 3 The corresponding description.

[0098] In one example, the training samples used to train the region of interest determination model are obtained through the following steps: acquiring a sample remote sensing image, wherein the sample remote sensing image displays a sample boundary region containing sample points of interest and the location of the sample points of interest; determining multiple boundary labeling points corresponding to the sample remote sensing image based on the intersection of a ray drawn from the origin of the polar coordinate system where the location of the sample points of interest is taken as the origin to the sample boundary region and the sample boundary region; and determining the sample remote sensing image and the corresponding multiple boundary labeling points as the training samples. The above operations can be referred to the foregoing. Figure 4 The corresponding descriptions in the embodiments.

[0099] The image region determination unit 1030 is configured to determine a region of interest (ROI) containing the target ROI based on the target number of ROI boundary prediction points. The operation of the image region determination unit 1030 can be referred to above. Figure 2 The operation described in section 230.

[0100] The region of interest determination unit 1040 is configured to determine a region of interest containing the target point of interest based on the region of interest to be determined in the region of interest image. The operation of the region of interest determination unit 1040 can be referred to above. Figure 2 The operation described in section 240.

[0101] Further reference Figure 11 , Figure 11 A block diagram of an example of an interest region determination unit 1100 in an apparatus for determining an interest region according to an embodiment of this specification is shown.

[0102] like Figure 11 As shown, the above-mentioned region of interest determination unit 1100 may include a road network undetermined region acquisition module 1110 and a region fusion module 1120.

[0103] The road network pending area acquisition module 1110 is configured to acquire a road network pending area containing the target point of interest. The aforementioned road network pending area can be determined based on road network information related to the target point of interest. The operation of the road network pending area acquisition module 1110 can be referred to above. Figure 5 The operation of 510 is described.

[0104] The region fusion module 1120 is configured to fuse the undetermined region of the region of interest image and the undetermined region of the road network of the region of interest to determine the region of interest containing the target point of interest. The operation of the region fusion module 1120 can be referred to above. Figure 5 The operation described in section 520.

[0105] Continue to refer to Figure 12 , Figure 12 A block diagram of an example of a region fusion module 1200 in an apparatus for determining a region of interest according to an embodiment of this specification is shown.

[0106] like Figure 12 As shown, the aforementioned region fusion module 1200 may include a distance determination submodule 1210, a distance determination submodule 1210, and a region determination submodule 1230.

[0107] The distance determination submodule 1210 is configured to determine, based on the location of the target point of interest, the distance from the target point of interest to at least one target location in the undetermined area of ​​the road network of the region of interest. The operation of the distance determination submodule 1210 can be referred to above. Figure 6 The process of determining the distance from a target point of interest to at least one target location in the undetermined area of ​​the road network of the region of interest.

[0108] In one example, the target location includes the vertices of the region of interest road network to be determined and / or the boundaries of the region of interest road network to be determined.

[0109] Boundary retention submodule 1220 is configured to determine the retention boundary of the undetermined area of ​​the road network in the region of interest based on at least one determined distance. The operation of boundary retention submodule 1220 can be referred to above. Figure 6 The process of determining the reserved boundaries is described.

[0110] In one example, the boundary preservation submodule 1220 described above can be further configured to: determine the minimum circumscribed circle radius of the region of interest (ROI) to be determined; determine the boundary of the ROI road network to be determined involving a distance not greater than the minimum circumscribed circle radius among at least one determined distance as the boundary to be preserved; and extract the preserved boundary of the ROI road network to be determined from the boundary to be preserved. The operation of the boundary preservation submodule 1220 described above can be referred to the above. Figure 7 The process of determining the reserved boundaries of the undetermined area of ​​the road network in the described region of interest 700.

[0111] The region determination submodule 1230 is configured to determine the region of interest containing the target point of interest based on the region to be determined in the region of interest image and the reserved boundary. The operation of the region determination submodule 1230 can be referred to above. Figure 8 The process of determining the region of interest as described.

[0112] In one example, the region determination submodule 1230 described above can be further configured to: determine the circumscribed polygon of the region of interest image to be determined for the non-reserved boundary of the undetermined boundary of the region of interest road network; and determine the region of interest containing the target point of interest based on the boundary of the circumscribed polygon and the reserved boundary. The operation of the region determination submodule 1230 can be referred to above. Figure 8 The process of determining the region of interest as described.

[0113] Continue to refer to Figure 13 , Figure 13 A block diagram of yet another example of a device 1300 for determining a region of interest according to an embodiment of this specification is shown.

[0114] like Figure 13 As shown, the apparatus 1300 for determining the region of interest may include an image acquisition unit 1310, a boundary point prediction unit 1320, an image region to be determined determination unit 1330, a region of interest determination unit 1340, a vector acquisition unit 1350, an accuracy determination unit 1360, and an information storage unit 1370.

[0115] The vector acquisition unit 1350 is configured to acquire feature vectors corresponding to the determined urban functional areas. The operation of the vector acquisition unit 1350 can be referred to above. Figure 9 The operation of 950 is described.

[0116] The accuracy determination unit 1360 is configured to provide the feature vector to the urban functional area assessment model to obtain the determination accuracy of the corresponding urban functional area. Figure 9 The operation described for 960.

[0117] The information entry unit 1370 is configured to store the urban functional area information into a preset database in response to the determination accuracy being greater than a predetermined threshold. The operation of the information entry unit 1370 can be referred to above. Figure 9 The operation of 970 is described.

[0118] It should be noted that the aforementioned region of interest may include urban functional areas. The image acquisition unit 1310, boundary point prediction unit 1320, image region to be determined unit 1330, and region of interest determination unit 1340 can be referenced. Figure 10 The corresponding descriptions of the image acquisition unit 1010, boundary point prediction unit 1020, image region determination unit 1030, and region of interest determination unit 1040 in the embodiment will not be repeated here.

[0119] Reference above Figures 1 to 13Embodiments of the method and apparatus for determining a region of interest according to embodiments of this specification have been described.

[0120] The apparatus for determining a region of interest in the embodiments of this specification can be implemented in hardware, software, or a combination of hardware and software. Taking software implementation as an example, as a logically defined apparatus, it is formed by the processor of the device loading corresponding computer program instructions from memory into RAM and executing them. In the embodiments of this specification, the apparatus for determining a region of interest can, for example, be implemented using an electronic device.

[0121] Figure 14 A schematic diagram of an apparatus 1400 for determining a region of interest according to an embodiment of this specification is shown.

[0122] like Figure 14 As shown, the apparatus 1400 for determining a region of interest may include at least one processor 1410, a memory (e.g., non-volatile memory) 1420, a memory 1430, and a communication interface 1440, and the at least one processor 1410, memory 1420, memory 1430, and communication interface 1440 are connected together via a bus 1450. At least one processor 1410 executes at least one computer-readable instruction (i.e., the elements implemented in software described above) stored or encoded in the memory.

[0123] In one embodiment, computer-executable instructions are stored in memory that, when executed, cause at least one processor 1410 to: acquire a remote sensing image containing a target point of interest; provide the remote sensing image containing the target point of interest to a region of interest determination model to obtain a target number of predicted boundary points of the region of interest; determine a region of interest image to be determined containing the target point of interest based on the target number of predicted boundary points of the region of interest; and determine a region of interest containing the target point of interest based on the region of interest image to be determined.

[0124] It should be understood that the computer-executable instructions stored in memory, when executed, cause at least one processor 1410 to perform the above-described combinations in the various embodiments of this specification. Figure 1-9 The description includes various operations and functions.

[0125] Specifically, a system or apparatus equipped with a readable storage medium may be provided, on which software program code implementing the functions of any of the embodiments described above is stored, and the computer or processor of the system or apparatus can read and execute the instructions stored in the readable storage medium.

[0126] In this case, the program code itself, which can be read from a readable medium, can perform the functions of any of the above embodiments. Therefore, the machine-readable code and the readable storage medium storing the machine-readable code constitute a part of the present invention.

[0127] The computer program code required for the operation of each part of this manual can be written in any one or more programming languages, including object-oriented programming languages ​​such as Java, Scala, Smalltalk, Eiffel, JADE, Emerald, C++, C#, VB, .NET, and Python; conventional procedural programming languages ​​such as C, Visual Basic 2003, Perl, COBOL 2002, PHP, and ABAP; dynamic programming languages ​​such as Python, Ruby, and Groovy; or other programming languages. This program code can run on the user's computer, or as a standalone software package on the user's computer, or partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the latter case, the remote computer can be connected to the user's computer via any network, such as a local area network (LAN) or wide area network (WAN), or connected to an external computer (e.g., via the Internet), or in a cloud computing environment, or used as a service, such as Software as a Service (SaaS).

[0128] Examples of readable storage media include floppy disks, hard disks, magneto-optical disks, optical disks (such as CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RAM, DVD-RW, DVD-RW), magnetic tapes, non-volatile memory cards, and ROMs. Alternatively, program code can be downloaded from a server computer or the cloud via a communication network.

[0129] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0130] Not all steps and units in the above process and system structure diagrams are mandatory; some steps or units can be omitted as needed. The execution order of each step is not fixed and can be determined as required. The device structure described in the above embodiments can be a physical structure or a logical structure. That is, some units may be implemented by the same physical entity, or some units may be implemented by multiple physical entities, or they may be jointly implemented by certain components in multiple independent devices.

[0131] The term "exemplary" as used throughout this specification means "serving as an example, instance, or illustration" and does not imply that it is "preferred" or "advantageous" over other embodiments. Detailed descriptions are included for the purpose of providing an understanding of the described techniques. However, these techniques may be practiced without these detailed descriptions. In some instances, well-known structures and apparatuses are shown in block diagram form to avoid obscuring the concepts of the described embodiments.

[0132] The optional embodiments of the present specification have been described in detail above with reference to the accompanying drawings. However, the embodiments of the present specification are not limited to the specific details in the above embodiments. Within the scope of the technical concept of the embodiments of the present specification, various simple modifications can be made to the technical solutions of the embodiments of the present specification, and these simple modifications all fall within the protection scope of the embodiments of the present specification.

[0133] The foregoing description of this specification is provided to enable any person skilled in the art to implement or use the content of this specification. Various modifications to the content of this specification will be apparent to those skilled in the art, and the general principles defined herein can be applied to other variations without departing from the scope of protection of this specification. Therefore, this specification is not limited to the examples and designs described herein, but is consistent with the widest scope of the principles and novel features disclosed herein.

Claims

1. A method for determining a region of interest, comprising: Acquire remote sensing images containing target points of interest; The remote sensing image containing the target interest points is provided to the region of interest determination model to obtain the target number of predicted boundary points of the region of interest. Based on the target number of region of interest boundary prediction points, determine the region of interest image to be determined that contains the target region of interest points; as well as Based on the undetermined region of interest in the region of interest image, a region of interest containing the target point of interest is determined, wherein the region of interest includes urban functional areas; Obtain feature vectors corresponding to the determined urban functional areas. The feature vectors are determined based on the feature information corresponding to the determined urban functional areas. The feature information is used to evaluate whether the determined urban functional areas are accurate. The feature information includes at least one of the following: geometric topology matching information, population activity distribution patterns, and address keyword matching information. as well as The feature vector is provided to the urban functional area assessment model to obtain the accuracy of the corresponding urban functional area determination.

2. The method as described in claim 1, wherein, The step of determining the region of interest containing the target point of interest based on the region of interest image to be determined includes: Obtain a potential road network region containing the target point of interest, wherein the potential road network region is determined based on road network information relating to the target point of interest; and The region of interest containing the target point of interest is determined by fusing the undetermined region of the region of interest image and the undetermined region of the road network in the region of interest.

3. The method as described in claim 2, wherein, The step of fusing the undetermined region of the region of interest image and the undetermined region of the road network to determine the region of interest containing the target point of interest includes: Based on the location of the target point of interest, determine the distance from the target point of interest to at least one target location in the undetermined area of ​​the road network of the interest region; Based on at least one determined distance, determine the reserved boundary of the undetermined area of ​​the road network in the region of interest; and Based on the undetermined region of interest in the region of interest image and the reserved boundary, the region of interest containing the target point of interest is determined.

4. The method of claim 3, wherein, The target location includes the vertices of the undetermined region of the region of interest road network and / or the boundaries of the undetermined region of the region of interest road network. The step of determining the reserved boundary of the undetermined area of ​​the road network of the region of interest based on at least one determined distance includes: Determine the minimum circumcircle radius of the region to be determined in the region of interest image; The boundary of the undetermined area of ​​the road network in the region of interest involving at least one determined distance that is not greater than the radius of the minimum circumcircle is determined as the boundary to be retained; and Extract the reserved boundaries of the undetermined area of ​​the road network in the region of interest from the boundary to be reserved.

5. The method of claim 3, wherein, The step of determining the region of interest containing the target interest point based on the region of interest to be determined in the region of interest image and the preserved boundary includes: For the non-preserved boundary of the undetermined region of the road network in the region of interest, determine the circumscribed polygon of the undetermined region of the image of the region of interest; and Based on the boundary of the circumscribed polygon and the reserved boundary, the region of interest containing the target point of interest is determined.

6. The method of claim 1, wherein, The training samples used to train the region of interest determination model are obtained through the following steps: Acquire sample remote sensing images, wherein the sample remote sensing images display sample boundary regions containing sample interest points and the locations of the sample interest points; Based on the intersection of the ray drawn from the origin of the polar coordinate system (where the sample point of interest is the origin) to the sample boundary region and the sample boundary region, multiple boundary annotation points corresponding to the sample remote sensing image are determined; and The sample remote sensing image and its corresponding multiple boundary annotation points are determined as the training sample.

7. The method as described in any one of claims 1 to 6, The method further includes: In response to the determination accuracy being greater than a predetermined threshold, the urban functional area information is stored in a preset database.

8. An apparatus for determining a region of interest, comprising: The image acquisition unit is configured to acquire remote sensing images containing target points of interest; The boundary point prediction unit is configured to provide the remote sensing image containing the target interest points to the region of interest determination model to obtain the target number of interest region boundary prediction points; The image region determination unit is configured to determine the region of interest in the image containing the target interest points based on the target number of region of interest boundary prediction points. as well as The region of interest determination unit is configured to determine a region of interest containing the target point of interest based on the region of interest to be determined region of the region of interest image, wherein the region of interest includes urban functional areas. The vector acquisition unit is configured to acquire a feature vector corresponding to the determined urban functional area. The feature vector is determined based on the acquired feature information corresponding to the determined urban functional area. The feature information is used to evaluate whether the determined urban functional area is accurate. The feature information includes at least one of the following: geometric topology matching information, population activity distribution pattern, and address keyword matching information. as well as An accuracy determination unit is configured to provide the feature vector to the urban functional area assessment model to obtain the determination accuracy of the corresponding urban functional area.

9. An apparatus for determining a region of interest, comprising: At least one processor, a memory coupled to the at least one processor, and a computer program stored on the memory, wherein the at least one processor executes the computer program to implement the method for determining a region of interest as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program that, when executed by a processor, implements the method for determining a region of interest as described in any one of claims 1 to 7.

11. A computer program product comprising a computer program executed by a processor to implement the method for determining a region of interest as described in any one of claims 1 to 7.

Images