A boundary determination method, device, storage medium and electronic device

Abstract

In the boundary determination method provided in the specification, a target area requiring division of an AOI is determined in advance; in determining the AOI boundary in the target area, a satellite image and a user distribution image of the target area are first acquired, and a road network image of the target area is determined according to the satellite image; the obtained satellite image, road network image and user distribution image of the target area are input into a pre-trained first semantic segmentation model to obtain a pending boundary image of the AOI in the target area, and the boundary of the AOI is determined according to the pending boundary image. When the boundary determination method provided in the specification is used to determine the boundary of the AOI, multi-source images can be acquired, and the boundary of the AOI is determined in combination with a pre-trained model; unlike the traditional manual boundary determination method, all steps in the method can be completed by an electronic device, which effectively saves human resources and greatly improves the efficiency of determining the boundary of the AOI.

Description

TECHNICAL FIELD

[0001] The present specification relates to the technical field of computer, and particularly relates to a boundary determination method and device, a storage medium and an electronic device. BACKGROUND

[0002] Nowadays, the area of interest (AOI) is widely used in many industries related to electronic maps. The area of interest is divided in the electronic map according to the social function, for example, the cell, school, shopping mall, office building and the like are divided on the map. Generally, the boundary of a region is determined to quickly and effectively divide the region. Therefore, in order to obtain a real and reliable region division, the method for determining the boundary of the region is crucial.

[0003] The traditional method divides the region by using the artificial way to determine the boundary of the region, such as collecting the actual road and satellite map data manually, manually drawing the boundary and the like.

[0004] However, there are many disadvantages in the process of manually determining the boundary of the region, such as the poor quality of the collected data, the long auditing period of the data, the high labor cost and the like. These problems lead to the low efficiency of the existing boundary determination method and the cumbersome implementation process. SUMMARY

[0005] The present specification provides a boundary determination method, device, storage medium and electronic device to at least partially solve the above problems in the prior art.

[0006] The present specification adopts the following technical solutions:

[0007] The present specification provides a boundary determination method, comprising:

[0008] According to the preset target region, a satellite image of the target region is obtained, according to the satellite image, a road network image of the target region is determined, and according to the waybill information in the target region, a user distribution image in the target region is determined;

[0009] The satellite image, the road network image and the user distribution image are input into a pre-trained first semantic segmentation model to obtain a pending boundary image of the area of interest (AOI) in the target region output by the first semantic segmentation model, and the pending boundary image contains a pending AOI boundary.

[0010] According to the pending boundary image, the boundary of the AOI is determined.

[0011] Optionally, according to the satellite image, the road network image of the target region is determined, specifically comprising:

[0012] determining an estimated boundary of an AOI in the target area according to the satellite image;

[0013] determining an estimated range of the AOI according to the estimated boundary;

[0014] enlarging the estimated range, and determining an enlarged range as an extended range;

[0015] obtaining road network information in the extended range;

[0016] determining a road network image of the target area according to the road network information.

[0017] Optionally, a user distribution image is obtained according to shipping order information in the target area, and specifically includes:

[0018] mapping each shipping order information in the target area to an image coordinate;

[0019] determining a user distribution image in the target area according to a distribution of the image coordinates in the target area.

[0020] Optionally, a boundary of the AOI is determined according to the to-be-determined boundary image, and specifically includes:

[0021] inputting the to-be-determined boundary image into a pre-trained second semantic segmentation model to correct a to-be-determined AOI boundary in the to-be-determined boundary image by the second semantic segmentation model, and obtaining a corrected boundary image containing a corrected AOI boundary;

[0022] determining the boundary of the AOI according to the corrected boundary image.

[0023] Optionally, a boundary of the AOI is determined according to the corrected boundary image, and specifically includes:

[0024] dilating the corrected AOI boundary in the corrected boundary image to obtain a dilated AOI boundary;

[0025] extracting the dilated AOI boundary to obtain an AOI boundary with a specified width of pixels;

[0026] mapping an image coordinate of each pixel point on the AOI boundary with the specified width of pixels to a world coordinate;

[0027] determining an AOI boundary in the target area according to the world coordinate.

[0028] Optionally, a first semantic segmentation model is pre-trained, and specifically includes:

[0029] According to a preset sample area, a satellite image of the sample area is obtained, and according to the satellite image, a road network image of the sample area is determined; according to waybill information in the sample area, a user distribution image in the sample area is determined;

[0030] The satellite image, the road network image and the user distribution image are input into a pre-trained first semantic segmentation model to obtain a to-be-optimized undetermined boundary image of an AOI in the sample area output by the first semantic segmentation model;

[0031] The first semantic segmentation model is trained with the minimum difference between the to-be-optimized undetermined boundary image and a real boundary image of the AOI in the sample area as an optimization target.

[0032] Optionally, a second semantic segmentation model is pre-trained, and the pre-training specifically includes:

[0033] A real boundary image of the AOI in the sample area is obtained;

[0034] The real boundary image is scrambled to obtain a scrambled boundary image containing a scrambled AOI boundary;

[0035] The scrambled boundary image is input into the second semantic segmentation model to correct the scrambled AOI boundary in the scrambled boundary image by the second semantic segmentation model, and a to-be-optimized corrected boundary image is obtained;

[0036] The second semantic segmentation model is trained with the minimum difference between the to-be-optimized corrected boundary image and the real boundary image of the AOI in the sample area as an optimization target.

[0037] The present specification provides a boundary determination device, and the device includes:

[0038] An image acquisition module obtains a satellite image of a target area according to a preset target area, and determines a road network image of the target area according to the satellite image; and a user distribution image in the target area is determined according to waybill information in the target area;

[0039] An undetermined image determination module inputs the satellite image, the road network image and the user distribution image into a pre-trained first semantic segmentation model to obtain an undetermined boundary image of an area of interest (AOI) in the target area output by the first semantic segmentation model, and the undetermined boundary image contains an undetermined AOI boundary;

[0040] A boundary determination module determines a boundary of the AOI according to the undetermined boundary image.

[0041] The specification provides a computer readable storage medium, the storage medium stores a computer program, the computer program is executed by a processor to realize the boundary determination method.

[0042] The specification provides an electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor executes the program to realize the boundary determination method.

[0043] The above at least one technical solution adopted by the specification can achieve the following beneficial effects:

[0044] In the boundary determination method provided by the specification, the target area requiring AOI division is determined in advance; when determining the AOI boundary in the target area, the satellite image and the user distribution image of the target area are first obtained, and the road network image of the target area is determined according to the satellite image; the obtained satellite image, road network image and user distribution image of the target area are input into the pre-trained first semantic segmentation model to obtain the undetermined boundary image of the AOI in the target area, and the boundary of the AOI is determined according to the undetermined boundary image. When determining the boundary of the AOI by using the boundary determination method provided by the specification, multiple source images can be obtained, and the boundary of the AOI is determined in combination with the pre-trained model; unlike the traditional manual boundary determination method, all the steps in the method can be completed by an electronic device, which effectively saves human resources and greatly improves the efficiency of determining the AOI boundary. BRIEF DESCRIPTION OF DRAWINGS

[0045] The drawings described herein are used to provide further understanding of the specification, and form a part of the specification. The illustrative embodiments of the specification and their descriptions serve to explain the specification, and do not constitute an improper limitation on the specification. In the drawings:

[0046] Figure 1 It is a flowchart of the boundary determination method in the specification;

[0047] Figure 2 It is a schematic diagram of the undetermined AOI boundary determined by the first semantic segmentation model in the specification;

[0048] Figure 3 It is a schematic diagram of the corrected AOI boundary determined by the second semantic segmentation model in the specification;

[0049] Figure 4 It is a schematic diagram of the dilated AOI boundary obtained by dilating the corrected AOI boundary in the specification;

[0050] Figure 5 It is a schematic diagram of the AOI boundary with a specified width obtained by extracting the dilated AOI boundary in the specification;

[0051] Figure 6 A schematic diagram of a boundary determination device provided for the present specification;

[0052] Figure 7 An electronic device schematic diagram corresponding to Figure 1 provided for the present specification. DETAILED DESCRIPTION

[0053] At present, the main method for determining boundaries in the industry when dividing AOI is manual determination of boundaries, and the main process includes manual data collection, manual review, manual drawing, etc. Among them, manual collection includes obtaining data from images and surveying on actual roads, etc. In fact, in order to more conveniently integrate and utilize the collected data, a set of standard collection rules will be preset when manually collecting data, and all collectors will collect data based on the rules.

[0054] However, due to individual differences, the data collected by the collectors may have some deviations. When the deviation is too large, the collected data will become invalid data, which directly leads to low overall quality and poor efficiency of manual data collection. At the same time, the difference between each set of data also increases the difficulty of manual review and reduces the review speed. In addition, since the division of AOI often requires a large number of personnel to participate, the labor cost is usually high.

[0055] In order to solve the above problems, the present specification provides a boundary determination method that can be implemented by an electronic device and can replace manual collection.

[0056] In order to make the purpose, technical scheme and advantages of the present specification clearer, the technical scheme of the present specification will be described clearly and completely in combination with the specific embodiments of the present specification and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present specification, not all. Based on the embodiments in the present specification, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of the present application.

[0057] The technical scheme provided by each embodiment of the present specification will be described in detail below in combination with the drawings.

[0058] Figure 1 A flowchart of a boundary determination method in the present specification, specifically comprising the following steps:

[0059] S100: Obtain a satellite image of a target region according to a preset target region, determine a road network image of the target region according to the satellite image, and determine a user distribution image in the target region according to shipping order information in the target region.

[0060] The image segmentation method provided in the specification can be completed by any electronic device with computing functions, including but not limited to terminals, servers, and the like.

[0061] In actual application, a target area is usually determined in advance, and the AOI in the target area is divided, i.e., the boundary of the AOI in the target area is determined. Therefore, before the method is executed, the target area in which the boundary of the AOI needs to be determined is first preset. The position, size, and the like of the preset target area can be set arbitrarily according to requirements, which are not limited in the specification.

[0062] In determining the boundary of the AOI in the target area, the multi-source image of the target area can be acquired first, which can include the satellite image of the target area, the road network image of the target area, the user distribution image in the target area, and the like, wherein the road network image can be obtained according to the satellite image. Specifically, for each road network in the satellite image, the road network in the road network image can be mapped to the width and pixel value corresponding to the road network level of the road network according to the road network level.

[0063] In addition, in order to avoid losing information in the process of acquiring the image, when the satellite image of the target area is acquired, the range of the target area can be expanded, and the satellite image in the expanded range can be acquired, so as to ensure that the range of the satellite image can completely cover the range of the target area and no information is lost.

[0064] S102: inputting the satellite image, the road network image, and the user distribution image into a pre-trained first semantic segmentation model to obtain the pending boundary image of the AOI in the target area output by the first semantic segmentation model, wherein the pending boundary image contains the pending AOI boundary.

[0065] The satellite image, the road network image, and the user distribution image of the target area obtained in step S100 are input into the pre-trained first semantic segmentation model to obtain the pending boundary image containing the pending boundary of the AOI in the target area output by the first semantic segmentation model. When each image is input, each image can be input into the first semantic segmentation model individually, or the layers of the images can be superimposed to form a multi-channel image and then input into the first semantic segmentation model.

[0066] The first semantic segmentation model used in this step can be any model capable of outputting an image containing the tentative AOI boundary within the target area. This specification provides an example here: the first semantic segmentation model can be a U-net model. After inputting the satellite image, road network image, and user distribution image within the target area into the U-net model, the model can obtain the probability of each pixel point in the output tentative boundary image belonging to the tentative AOI boundary according to the pixel difference of the satellite image, road network shape, and user distribution, and can determine the pixel points with a probability greater than a specified threshold as the pixel points on the tentative AOI boundary, and set the pixel value of the pixel points to 1, and otherwise, set the pixel value of the pixel points not on the tentative AOI boundary to 0, to obtain a binary image about the tentative AOI boundary, i.e., the tentative boundary image.

[0067] In addition to the U-net model described above, other models can also be used as the first semantic segmentation model to obtain the tentative boundary image, which will not be described in detail here.

[0068] S104: determining the boundary of the AOI according to the tentative boundary image.

[0069] According to the tentative boundary image containing the tentative AOI boundary obtained in step S104, the boundary of the AOI within the target area is determined.

[0070] When the boundary determination method provided by the specification is used to determine the AOI boundary, the electronic device can collect multi-source images of the target area, i.e., satellite images, road network images, and user distribution images, and according to the collected multi-source images, a first semantic segmentation model pre-trained to obtain a tentative boundary image containing the tentative AOI boundary, and determine the boundary of the AOI according to the tentative boundary image. As can be seen, when implementing the boundary determination method provided by the specification, no manual operation is required at all, and all steps can be completed by the electronic device itself, which avoids the shortcomings of poor data quality, long data review period, and high labor cost, and greatly improves the efficiency of determining the AOI boundary.

[0071] In addition, the traditional method often fails to determine the AOI boundary due to poor road network quality when determining the boundary of the AOI. However, the method of determining the AOI boundary according to multi-source images in this specification can completely solve this problem. By determining the AOI boundary according to multiple different images, when one of the images has quality problems, the AOI boundary can still be determined according to the other images, effectively solving the problem of difficulty in determining the AOI boundary due to poor road network quality.

[0072] In step S100, a detailed method of obtaining a satellite image of a target region is described. Similarly, a road network image of the target region can also be obtained from a satellite image using a similar method. Specifically, an estimated boundary of an AOI in the target region can be determined from the satellite image; an estimated range of the AOI can be determined based on the estimated boundary; the estimated range can be expanded to determine an expanded range; road network information in the expanded range can be obtained; and a road network image of the target region can be determined based on the road network information.

[0073] The estimated boundary of the AOI in the target region is the approximate boundary of the AOI whose boundary needs to be determined, which is directly drawn from the satellite image. When the estimated range is expanded, the size of the expanded range can be determined according to actual needs. The road network information can be obtained from the recorded road network data. For example, the map range corresponding to the expanded range can be found on the electronic map according to the expanded range, and the road network information in the expanded range can be obtained according to the road network annotations on the map.

[0074] Meanwhile, in step S100, there are various methods of obtaining a user distribution image based on the shipping order information in the target region. Specifically, for each shipping order information in the target region, the shipping order information can be mapped to an image coordinate; and a user distribution image in the target region can be determined based on the distribution of the image coordinates in the target region.

[0075] The shipping order information can be obtained by various methods, such as shipping order information of instant delivery services or online car-hailing services. When the shipping order information is mapped to an image coordinate, the world coordinate of the user's location can be directly converted to an image coordinate, and a pixel value can be added at the corresponding coordinate in the image to represent the user distribution; or the user's location can be converted to a string by geohash, which represents a specific positioning area in the image. Different strings represent different positioning areas, and adjacent user locations can be converted to the same string. When all the shipping order information in the target region is converted to a string, the number of shipping order information falling into each positioning area can be determined to determine the pixel value of the positioning area, so as to represent the user distribution. In other words, the user distribution image can be regarded as a heat image. In the target region, the more users fall in a location, the higher the heat of the location, and the larger the pixel value. The user distribution image can be obtained based on the heat distribution in the target region.

[0076] In actual applications, any model is difficult to output completely accurate results. In step S102, the undetermined AOI boundary in the undetermined boundary image output by the first semantic segmentation model may have burrs due to obstacles in the input image, or the boundary may be disconnected due to missing of some pixel points, such asFigure 2 As shown, therefore, before obtaining the boundary of the AOI from the undetermined boundary image, the undetermined AOI boundary in the undetermined boundary image can be corrected first. Specifically, the undetermined boundary image can be input into a pre-trained second semantic segmentation model to correct the undetermined AOI boundary in the undetermined boundary image through the second semantic segmentation model, resulting in a boundary containing, as shown in the image. Figure 3 The image shows a corrected boundary image of the AOI boundary; the boundary of the AOI is determined based on the corrected boundary image. The resulting corrected boundary image is a binary image about the corrected AOI boundary, that is, the pixel value of the pixels on the corrected AOI boundary is 1, and the pixel value of other pixels is 0.

[0077] Similarly, the second semantic segmentation model used in the above method can also be any model that can correct the undetermined boundaries of AOI, such as the U-net model, etc., which will not be elaborated here.

[0078] When the second semantic segmentation model is used to correct the boundary of the AOI to be determined, and a corrected boundary image containing the corrected AOI boundary is obtained, the boundary of the AOI can be determined based on the corrected boundary image. However, the second semantic segmentation model also cannot guarantee that the boundary of the AOI to be determined can be perfectly corrected into a corrected AOI boundary without any loss of information. That is to say, there may still be some missing pixels in the corrected AOI boundary. Although only a few pixels are missing and will not cause the boundary to be broken, it may still affect the boundary of the final obtained AOI. Therefore, further optimization can be made to the corrected AOI boundary to ensure that a perfect AOI boundary can be obtained.

[0079] Specifically, the corrected AOI boundary in the corrected boundary image can be dilated to obtain, as shown below. Figure 4 The expanded AOI boundary is shown; the expanded AOI boundary is extracted to obtain an AOI boundary with a pixel width of a specified width; for each pixel on the AOI boundary with a pixel width of the specified width, the image coordinates of the pixel are mapped to world coordinates; the AOI boundary in the target area is determined based on the world coordinates.

[0080] By dilating the corrected AOI boundary, all points near the corrected AOI boundary are filled, ensuring that the resulting dilated AOI boundary is a connected region without any gaps. Subsequently, the frame of the dilated AOI boundary can be extracted, that is, the AOI boundary with a specified pixel width can be extracted, such as... Figure 5As shown. Among them, the specified width can be set according to the needs, and can be set to 1. At this time, the image coordinates of each pixel point on the AOI boundary of the specified width have a unique corresponding world coordinate. After mapping the image coordinates of each pixel point to the world coordinate, the boundary of the AOI in the target area can be determined according to the obtained world coordinate. Among them, the world coordinate can be the actual latitude and longitude.

[0081] In addition, the first semantic segmentation model used in the specification is a pre-trained model, and the training method can be specifically: obtaining a satellite image of a sample area according to a preset sample area; determining a road network image of the sample area according to the satellite image; determining a user distribution image in the sample area according to the waybill information in the sample area; inputting the satellite image, the road network image and the user distribution image into the pre-trained first semantic segmentation model to obtain an AOI boundary image to be optimized in the sample area output by the first semantic segmentation model; and training the first semantic segmentation model with the minimum difference between the AOI boundary image to be optimized and the real boundary image of the AOI in the sample area as the optimization target.

[0082] Similarly, there is also a pre-trained second semantic segmentation model in the specification, and the training method of the model can be specifically: obtaining a real boundary image of an AOI in a sample area; scrambling the real boundary image to obtain a scrambled boundary image containing a scrambled AOI boundary; inputting the scrambled boundary image into the second semantic segmentation model to correct the scrambled AOI boundary in the scrambled boundary image through the second semantic segmentation model to obtain an optimized correction boundary image; and training the second semantic segmentation model with the minimum difference between the optimized correction boundary image and the real boundary image of the AOI in the sample area as the optimization target. Among them, the scrambling of the real boundary image can be realized by random mask, pixel value change and the like.

[0083] It should be noted that in the above method of pre-training the first semantic segmentation model and the second semantic segmentation model, the sample area is an area whose AOI boundary has been determined in history, so the already determined AOI boundary corresponding to the sample area can be used as the real AOI boundary, and the image containing the real AOI boundary, that is, the real boundary image, can be used as the training label.

[0084] The above is the boundary determination method provided by the specification. Based on the same idea, the specification also provides a corresponding boundary determination device, such as Figure 6 As shown.

[0085] Figure 6 A boundary determination device provided by the specification is shown in the schematic diagram, which specifically comprises:

[0086] The image acquisition module 200 obtains a satellite image of the target area according to a preset target area, determines a road network image of the target area according to the satellite image, and determines a user distribution image in the target area according to waybill information in the target area.

[0087] The pending image determination module 202 inputs the satellite image, the road network image, and the user distribution image into a pre-trained first semantic segmentation model to obtain a pending boundary image of an area of interest (AOI) in the target area output by the first semantic segmentation model, wherein the pending boundary image contains a pending AOI boundary.

[0088] The boundary determination module 204 determines the boundary of the AOI according to the pending boundary image.

[0089] In an optional embodiment:

[0090] The image acquisition module 200 is specifically configured to determine an estimated boundary of an AOI in the target area according to the satellite image, determine an estimated range of the AOI according to the estimated boundary, expand the estimated range, determine an expanded range after expansion as an augmented range, acquire road network information in the augmented range, and determine a road network image of the target area according to the road network information.

[0091] In an optional embodiment:

[0092] The image acquisition module 200 is specifically configured to map each waybill information in the target area to an image coordinate, and determine a user distribution image in the target area according to a distribution of the image coordinates in the target area.

[0093] In an optional embodiment:

[0094] The boundary determination module 204 is specifically configured to input the pending boundary image into a pre-trained second semantic segmentation model to correct the pending AOI boundary in the pending boundary image through the second semantic segmentation model, obtain a corrected boundary image containing a corrected AOI boundary, and determine the boundary of the AOI according to the corrected boundary image.

[0095] In an optional embodiment:

[0096] The boundary determination module 204 is specifically configured to dilate the corrected AOI boundary in the corrected boundary image to obtain a dilated AOI boundary; extract the dilated AOI boundary to obtain an AOI boundary with a pixel width of a specified width; map image coordinates of each pixel point on the AOI boundary with the pixel width of the specified width to world coordinates; and determine the AOI boundary in the target region according to the world coordinates.

[0097] In an optional embodiment,

[0098] The device further includes a first training module 206, which is specifically configured to obtain a satellite image of a sample region according to a preset sample region, determine a road network image of the sample region according to the satellite image, determine a user distribution image in the sample region according to shipment information in the sample region, input the satellite image, the road network image, and the user distribution image into a pre-trained first semantic segmentation model to obtain a to-be-optimized undetermined boundary image of an AOI in the sample region output by the first semantic segmentation model, and train the first semantic segmentation model with the minimum difference between the to-be-optimized undetermined boundary image and a real boundary image of the AOI in the sample region as an optimization target.

[0099] In an optional embodiment,

[0100] The device further includes a second training module 208, which is specifically configured to obtain a real boundary image of an AOI in a sample region, scramble the real boundary image to obtain a scrambled boundary image containing scrambled AOI boundaries, input the scrambled boundary image into a second semantic segmentation model to correct the scrambled AOI boundaries in the scrambled boundary image by the second semantic segmentation model to obtain a to-be-optimized corrected boundary image, and train the second semantic segmentation model with the minimum difference between the to-be-optimized corrected boundary image and the real boundary image of the AOI in the sample region as an optimization target.

[0101] The present specification also provides a computer-readable storage medium storing a computer program, which can be used to execute the above Figure 1 The boundary determination method is provided.

[0102] The present specification also provides Figure 7 The schematic structural diagram of the electronic device is shown. As shown in the Figure 7 At the hardware level, the electronic device includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory, and of course, other hardware required by the business. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs to implement the above Figure 1The boundary determination method. Of course, in addition to the software implementation, the present specification does not exclude other implementations, such as logic devices or a combination of software and hardware, and the like, that is, the execution subject of the following processing flow is not limited to each logic unit, but can also be hardware or a logic device.

[0103] It should be noted that all the actions of obtaining signals, information or data in this application are carried out in accordance with the corresponding data protection regulations and policies of the country where the device is located, and with the authorization given by the owner of the corresponding device.

[0104] In the 1990s, it was quite obvious to distinguish whether an improvement in a technology was in hardware (e.g., improvement in circuit structures of diodes, transistors, switches, etc.) or in software (improvement in method flow). However, as technology has evolved, many improvements in method flow today can be considered as direct improvements in hardware circuit structures. Designers almost always obtain the corresponding hardware circuit structures by programming the improved method flow into hardware circuits. Therefore, it cannot be said that an improvement in a method flow cannot be implemented by hardware entity modules. For example, a programmable logic device (PLD) (e.g., a field programmable gate array (FPGA)) is an integrated circuit whose logic function is determined by user programming of the device. A digital system is "integrated" on a PLD by the designer programming it, rather than by asking a chip manufacturer to design and fabricate a custom integrated circuit chip. Moreover, instead of manually fabricating integrated circuit chips, this programming is now mostly implemented by "logic compiler" software, which is similar to software compilers used in program development, and the original code to be compiled is written in a specific programming language, which is called a hardware description language (HDL), and there are many such languages, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, RHDL (Ruby Hardware Description Language), etc., and the most commonly used are VHDL (Very-High-Speed Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should be aware that, as long as the method flow is logically programmed in the above-mentioned hardware description languages and programmed into an integrated circuit, a hardware circuit implementing the logical method flow can be easily obtained.

[0105] The controller can be implemented in any suitable way, for example, the controller can take the form of a microprocessor or processor and a computer readable medium storing computer readable program code, such as software or firmware, executable by the (micro)processor, logic gates, switches, an application specific integrated circuit (ASIC), a programmable logic controller and an embedded microcontroller, examples of which include but are not limited to the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20 and Silicone Labs C8051F320, the memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art will also know that, in addition to being implemented in pure computer readable program code, the controller can equally well be implemented to perform the same functions using logic gates, switches, an application specific integrated circuit, a programmable logic controller and an embedded microcontroller, etc. by means of a logical programming of the method steps. The controller can thus be considered as a hardware component, and the means comprised therein for performing the various functions can be considered as structures within the hardware component. Alternatively, the means for performing the various functions can even be considered as both a software module implementing the method and a structure within the hardware component.

[0106] The systems, apparatuses, modules or units illustrated by the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer can be a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0107] For the sake of description, the above apparatuses are described in various units by functions respectively. Of course, the functions of each unit can be implemented in the same or multiple software and / or hardware in implementing the present specification.

[0108] Those skilled in the art will understand that the embodiments of the present application can be provided as a method, a system or a computer program product. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer usable program code.

[0109] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flow or blocks Figure 1 means for functionally implementing the steps listed in the flowchart block or blocks.

[0110] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the function specified in the flowchart block or blocks. Figure 1 one or more flow or blocks Figure 1 means for functionally implementing the steps listed in the flowchart block or blocks.

[0111] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flow or blocks ​ means for functionally implementing the steps listed in the flowchart block or blocks.

[0112] In one typical configuration, the computing device includes one or more processors (CPU's), input / output interfaces, network interfaces, and memory.

[0113] The memory can include non-persistent memory and / or persistent memory, such as flash memory, read-only memory (ROM), and / or volatile or non-volatile random access memory (RAM), among others. The memory is an example of computer-readable media.

[0114] Computer-readable media includes permanent and non-permanent, movable and non-movable media that can be implemented by any method or technology to store information. The information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information accessible to a computing device. According to the definition herein, computer-readable media does not include transitory media such as modulated data signals and carriers.

[0115] It should also be noted that the terms "comprising", "containing", or any other variant thereof are intended to cover non-exclusive inclusion, such that a process, method, article or apparatus that comprises a list of elements does not only include those elements, but can also include other elements not expressly listed or inherent to such process, method, article or apparatus. Without more limitations, the element defined by the statement "comprising a" does not exclude the presence of additional identical elements in the process, method, article or apparatus that includes the element.

[0116] Those skilled in the art will appreciate that embodiments of the present specification can be provided as methods, systems or computer program products. Therefore, the present specification can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Moreover, the present specification can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0117] The present specification can be described in the general context of computer-executable instructions, such as program modules, executed by computers. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The present specification can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in both local and remote computer storage media including storage devices.

[0118] The various embodiments described in this specification are described using a numbering of embodiments approach: these are each individually integrated contributions pertaining to different but related aspects of the description. Each of the various embodiments can stand on its own, and each can be combined with the subject matter of other embodiments to produce further embodiments. Where appropriate, therefore, the contents of the specification can be regarded as incorporating text of the detailed description under the provisions of 35 U.S.C. § 112(f), including the equivalents of any amended claims.

[0119] The above description is embodied in the form of embodiments only and is not intended to limit the present specification. The present specification can be variously changed and modified by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present specification should be included in the scope of the claims of the present application.

Claims

1. A boundary determination method, characterized in that, include: A satellite image of the target area is obtained based on a preset target area, and a road network image of the target area is determined based on the satellite image. Based on the waybill information within the target area, a user distribution image within the target area is determined; The satellite image, the road network image, and the user distribution image are input into a pre-trained first semantic segmentation model to obtain the undetermined boundary image of the AOI within the target area output by the first semantic segmentation model. The undetermined boundary image contains the undetermined AOI boundary. The boundary of the AOI is determined based on the undetermined boundary image, wherein the undetermined AOI boundary in the undetermined boundary image is corrected to obtain a corrected boundary image. The corrected AOI boundary in the corrected boundary image is dilated with the constraint that the obtained dilated AOI boundary has no empty connected components, to obtain a dilated AOI boundary. The dilated AOI boundary is extracted to obtain an AOI boundary with a pixel width of a specified width. For each pixel on the AOI boundary with a pixel width of the specified width, the image coordinates of the pixel are mapped to world coordinates. The AOI boundary in the target area is determined based on the world coordinates.

2. The method as described in claim 1, characterized in that, Based on the satellite imagery, the road network image of the target area is determined, specifically including: The estimated boundary of the AOI within the target area is determined based on the satellite imagery; Based on the predicted boundaries, the predicted range of the AOI is determined; The estimated range is expanded, and the expanded range is defined as the extended range. Obtain road network information within the expanded area; The road network image of the target area is determined based on the road network information.

3. The method as described in claim 1, characterized in that, Based on the waybill information within the target area, a user distribution image is obtained, specifically including: For each waybill information within the target area, the waybill information is mapped to image coordinates; Based on the distribution of image coordinates in the target area, determine the user distribution image within the target area.

4. The method as described in claim 1, characterized in that, Determining the boundary of the AOI based on the undetermined boundary image specifically includes: The undetermined boundary image is input into a pre-trained second semantic segmentation model to correct the undetermined AOI boundary in the undetermined boundary image through the second semantic segmentation model, so as to obtain a corrected boundary image containing the corrected AOI boundary. The boundary of the AOI is determined based on the corrected boundary image.

5. The method as described in claim 1, characterized in that, Pre-training the first semantic segmentation model includes: Satellite images of the sample area are obtained based on a preset sample area; road network images of the sample area are determined based on the satellite images; and user distribution images of the sample area are determined based on waybill information within the sample area. The satellite image, the road network image, and the user distribution image are input into a pre-trained first semantic segmentation model to obtain the undetermined boundary image of the AOI within the sample region output by the first semantic segmentation model. The first semantic segmentation model is trained with the goal of minimizing the difference between the undetermined boundary image to be optimized and the true boundary image of the AOI within the sample region.

6. The method as described in claim 4, characterized in that, Pre-training the second semantic segmentation model specifically includes: Obtain the true boundary image of the AOI within the sample area; The real boundary image is scrambled to obtain a scrambled boundary image containing the scrambled AOI boundary; The scrambled boundary image is input into the second semantic segmentation model to correct the scrambled AOI boundary in the scrambled boundary image, thereby obtaining the boundary image to be optimized and corrected. The second semantic segmentation model is trained with the goal of minimizing the difference between the boundary image to be optimized and the true boundary image of the AOI in the sample region.

7. A boundary determination device, characterized in that, include: The image acquisition module obtains satellite images of the target area based on a preset target area, and determines the road network image of the target area based on the satellite images; Based on the waybill information within the target area, a user distribution image within the target area is determined; The pending image determination module inputs the satellite image, the road network image, and the user distribution image into a pre-trained first semantic segmentation model to obtain the pending boundary image of the interest surface (AOI) within the target area output by the first semantic segmentation model, wherein the pending boundary image contains the pending AOI boundary. The boundary determination module determines the boundary of the AOI based on the undetermined boundary image. Specifically, it corrects the undetermined AOI boundary in the undetermined boundary image to obtain a corrected boundary image. Using the constraint that the obtained expanded AOI boundary has no empty connected components, it expands the corrected AOI boundary in the corrected boundary image to obtain an expanded AOI boundary. It then extracts the expanded AOI boundary to obtain an AOI boundary with a specified pixel width. For each pixel on the AOI boundary with the specified pixel width, it maps the image coordinates of that pixel to world coordinates and determines the AOI boundary within the target area based on the world coordinates.

8. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, implements the method described in any one of claims 1 to 6.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the method described in any one of claims 1 to 6.

Images