CN115527028A

CN115527028A - Map data processing method and device

Info

Publication number: CN115527028A
Application number: CN202210983182.3A
Authority: CN
Inventors: 逯飞; 吴彬; 钟开; 杨建忠; 张通滨; 卢振
Original assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Current assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Priority date: 2022-08-16
Filing date: 2022-08-16
Publication date: 2022-12-27

Abstract

The disclosure provides a map data processing method and device, relates to the field of data processing, and particularly relates to the field of car networking and intelligent cabins. The specific implementation scheme is as follows: and acquiring a remote sensing image of the first area. And performing first semantic segmentation processing according to the remote sensing image to obtain at least one first segmentation area. And performing second semantic segmentation processing on the remote sensing image according to the at least one first segmentation region to obtain at least one second segmentation region, wherein the processing precision of the second semantic segmentation processing is higher than that of the first semantic segmentation processing. And updating the area of interest (AOI) in the map data corresponding to the first area according to the at least one second divided area. The technical scheme of the present disclosure can effectively improve the efficiency and accuracy of updating the AOI region.

Description

Map data processing method and device

Technical Field

The disclosure relates to the field of car networking and intelligent cabins in data processing, in particular to a map data processing method and device.

Background

AOI (Area Of Interest) refers to a geographical entity in the form Of an Area in map data. Such as a residential area, a university, an office building, etc. For AOI areas in map data, updating is usually required to ensure the accuracy of the map data.

In order to quickly and accurately update the AOI area in the map data, it is necessary to provide an efficient and highly accurate map data processing method.

Disclosure of Invention

The disclosure provides a map data processing method and device.

According to a first aspect of the present disclosure, there is provided a map data processing method, including:

acquiring a remote sensing image of a first area;

performing first semantic segmentation processing according to the remote sensing image to obtain at least one first segmentation region;

according to the at least one first segmentation region, performing second semantic segmentation processing on the remote sensing image to obtain at least one second segmentation region, wherein the processing precision of the second semantic segmentation processing is higher than that of the first semantic segmentation processing;

and updating the AOI area of the interest surface in the map data corresponding to the first area according to the at least one second segmentation area.

According to a second aspect of the present disclosure, there is provided a map data processing apparatus including:

the acquisition module is used for acquiring a remote sensing image of the first area;

the first processing module is used for carrying out first semantic segmentation processing according to the remote sensing image to obtain at least one first segmentation area;

the second processing module is used for performing second semantic segmentation processing on the remote sensing image according to the at least one first segmentation region to obtain at least one second segmentation region, wherein the processing precision of the second semantic segmentation processing is higher than that of the first semantic segmentation processing;

and the updating module is used for updating the interest area AOI area in the map data corresponding to the first area according to the at least one second division area.

According to a third aspect of the present disclosure, there is provided an electronic device comprising:

at least one processor; and

a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of the first aspect.

According to a fourth aspect of the present disclosure, there is provided a non-transitory computer readable storage medium having stored thereon computer instructions for causing the computer to perform the method of the first aspect.

According to a fifth aspect of the present disclosure, there is provided a computer program product comprising: a computer program, stored in a readable storage medium, from which at least one processor of an electronic device can read the computer program, execution of the computer program by the at least one processor causing the electronic device to perform the method of the first aspect.

Techniques according to the present disclosure improve the efficiency and accuracy of updating AOI in map data.

It should be understood that the statements in this section do not necessarily identify key or critical features of the embodiments of the present disclosure, nor do they limit the scope of the present disclosure. Other features of the present disclosure will become readily apparent from the following description.

Drawings

The drawings are included to provide a better understanding of the present solution and are not to be construed as limiting the present disclosure. Wherein:

FIG. 1 is a schematic diagram of an AOI region provided by an embodiment of the present disclosure;

FIG. 2 is a schematic diagram illustrating an implementation of a remote sensing image provided by an embodiment of the present disclosure;

fig. 3 is a flowchart of a map data processing method provided in an embodiment of the present disclosure;

fig. 4 is a second flowchart of a map data processing method according to an embodiment of the present disclosure;

fig. 5 is a schematic diagram of an implementation of determining a first segmentation area according to an embodiment of the present disclosure;

fig. 6 is an implementation schematic diagram of determining a mapping region according to an embodiment of the present disclosure;

fig. 7 is a schematic diagram of an implementation of determining a second segmentation region according to an embodiment of the present disclosure;

fig. 8 is a schematic structural diagram of a map data processing apparatus according to an embodiment of the present disclosure;

fig. 9 is a block diagram of an electronic device for implementing a map data processing method according to an embodiment of the present disclosure.

Detailed Description

Exemplary embodiments of the present disclosure are described below with reference to the accompanying drawings, in which various details of the embodiments of the disclosure are included to assist understanding, and which are to be considered as merely exemplary. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the present disclosure. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.

In order to better understand the technical solution of the present disclosure, the related technologies related to the present disclosure are further described in detail below.

With the development of science and technology and the improvement of living standard of people, the demand of people for electronic maps is higher and higher. By taking Area Of Interest (AOI) data as a carrier, all production elements such as business data, production materials, production tools, operation strategies, people and the like Of an organization are organized, and online and offline integration can be effectively completed and the evolution Of the organization to an information space is promoted.

The AOI data is first briefly described below:

the Point of Information (POI) data in the map data represents an entity that actually exists in a physical world, such as a house, a restaurant, a building, a bus station, and the like.

For some specific POI data, there is corresponding fence information in the physical world, such as a cell, and there is a fenced area corresponding to the cell. The Area with the fence is used as an Area Of Interest (AOI for short), and can also be translated into an Interest plane.

AOI is a regional geographic entity in map data, such as a residential area, a college, an office building, an industrial park, a complex, a hospital, a scenic spot or a stadium, etc.

The AOI in the map usually includes four items of basic information, namely, name, address, category, and longitude and latitude. AOI has better computational power and better stability than POI.

For example, the AOI can be understood with reference to fig. 1, and fig. 1 is a schematic diagram of an AOI region provided in an embodiment of the present disclosure.

As shown in fig. 1, assuming a park currently exists, an AOI may be determined for the park. As can be understood from fig. 1, the area indicated by 101 is the fence area of park number one.

Four kinds of information can be included in the AOI corresponding to park one, wherein the name can be "park one" shown in fig. 1, and the address can be xx street xx in xx district xx city xx in xx provinces. And the category may be "park", for example, and the latitude and longitude may be the latitude and longitude of the center position of the area, for example, or may also be the latitude and longitude of each position point on the boundary of the area.

In an actual implementation process, for one AOI, the specific implementation of the four items of information included in the AOI may be set according to actual requirements, which is not limited in this embodiment.

It can be understood that, since there is a possibility that geographic entities in an actual scene may change, for example, an added geographic entity, a removed geographic entity, or a geographic entity with a changed fence area may occur, for the changed geographic entities, corresponding updates need to be performed in the map data to ensure correctness of AOI information in the map data.

Currently, in the prior art, when the AOI information is updated for the map data, a worker usually performs manual analysis on the collected image data to determine the change of the geographic entity, and then manually updates the change of the geographic entity into the AOI information in the map data.

However, the data size of the AOI information in the map data is relatively large, and this implementation manner of manual analysis may result in inefficient update of the AOI information in the map data.

Meanwhile, the premise that the staff can realize manual analysis is that the latest image data needs to be collected in a specific area.

The currently acquired image data is generally a common RGB image. On one hand, the common image has no accurate coordinate information, and the band is only RGB, so the common image can provide less image information. On the other hand, since the shooting angle of a general image is usually shot in a flat view relative to a geographic entity, and therefore it is difficult to provide full-view information of a specific area, a large number of images are usually required to be shot for a geographic entity to be provided for a worker to analyze, which results in high cost of updating the AOI and low updating efficiency. Based on this, in terms of the amount of information provided and the required cost, efficiency, etc., in the ordinary image, the ordinary RGB image has difficulty in satisfying the requirements of AOI information update.

Based on the problems in the prior art, the present disclosure provides a technical concept of performing AOI region update based on a remote sensing image, and the remote sensing image is understood with reference to fig. 2, and fig. 2 is a schematic diagram of implementing the remote sensing image provided by an embodiment of the present disclosure.

The remote sensing image is a film or a photo recording electromagnetic waves of various ground objects. Compared with a common image, the remote sensing image contains accurate geographical position information, and a scene under a real geographical coordinate can be reflected through the image; the spectral dimension of the remote sensing image is more than that of the common image, the common image only has information of three wave bands of RGB, the remote sensing image can use more wave bands, and the information of the remote sensing image is more comprehensive due to abundant spectral characteristics.

Therefore, compared with the common image, the remote sensing image can accurately record the electromagnetic wave information of the ground, provide more accurate geographical position information, and provide more abundant wave bands, so that the remote sensing image can provide more abundant ground information. Then more accurate AOI region updates can be achieved based on the telepresence images.

And, it can be confirmed in conjunction with fig. 2 that the remote sensing image is usually generated by a downward shot, so that it is not necessary to acquire a plurality of images for the same geographic entity, and the image acquisition cost can be effectively reduced.

On the basis of the related contents of the remote sensing image described above, the following further introduces the technical ideas of the present disclosure: based on the characteristics of the remote sensing image, semantic segmentation can be carried out based on the remote sensing image in the method, so that a plurality of areas are divided, then the AOI areas in the map data are updated according to the divided areas, and therefore the updating of the AOI areas in the map data can be rapidly and accurately achieved.

Based on the above description, the method for processing map data provided by the present disclosure is described below with reference to specific embodiments. It should be noted that the execution main body in each embodiment of the present disclosure may be a device with a data processing function, such as a local server, a cloud server, a processor, a chip, and the like, a specific selection of the execution main body is not limited in this embodiment, and the execution main body may be implemented according to an actual requirement, and all devices with a data processing function may be used as the execution main body in each embodiment of the present disclosure.

First, a map data processing method provided by the present disclosure is described with reference to fig. 3, and fig. 3 is a flowchart of the map data processing method provided by the embodiment of the present disclosure.

As shown in fig. 3, the method includes:

s301, obtaining a remote sensing image of the first area.

In this embodiment, the first area may be understood as an area that needs AOI update, and the specific division of the first area may be selected according to actual needs, for example, a certain administrative area may be used as the first area, or a certain area framed in a map may also be used as the first area.

After determining that the first region of AOI data needs to be updated, a telepresence image of the first region may be acquired.

In one possible implementation, the remote sensing images for the respective regions may be acquired in real time for AOI data update.

Alternatively, the remote sensing image may be acquired in advance and stored in a database, so that, for example, the remote sensing image of the first region may be directly acquired from the database. In this case, it is necessary to ensure that the remote sensing image stored in the database is the most recently acquired remote sensing image, so as to ensure the accuracy of the update of the map data.

And the remote sensing image may be, for example, an aerial image, that is, taken by an airplane. Alternatively, the remote sensing image may be, for example, a satellite image, that is, an image taken by a satellite, which is not limited in this embodiment.

S302, performing first semantic segmentation processing according to the remote sensing image to obtain at least one first segmentation area.

After the remote sensing image of the first area is obtained, semantic segmentation processing can be carried out on the remote sensing image. The semantic segmentation process is actually to classify images from pixel levels, and then classify pixels belonging to the same class into one class, thereby obtaining a plurality of segmented regions. Wherein one divided region obtained by division corresponds to a certain category.

Based on the above description, it can be determined that a remote sensing image includes many geographic entities, that is, the remote sensing image has many elements. And the remote sensing image has the advantages that the visual angle is a overlook, the distinction among different objects is small, and the characteristic of high local feature similarity is shown.

Therefore, if the remote sensing image is directly subjected to semantic segmentation, the obtained semantic segmentation result is relatively poor in effect. To solve this problem, in the present embodiment, two segmentation processes are performed on the remote sensing image.

First semantic segmentation processing can be performed according to the remote sensing image to obtain at least one first segmentation region.

S303, performing second semantic segmentation processing on the remote sensing image according to the at least one first segmentation region to obtain at least one second segmentation region, wherein the processing precision of the second semantic segmentation processing is higher than that of the first semantic segmentation processing.

And after the at least one first segmentation region is obtained, performing second semantic segmentation processing on the remote sensing image according to the at least one first segmentation region. In one possible implementation manner, for example, the second semantic segmentation processing may be performed on each first segmented region, so as to obtain at least one second segmented region.

In this embodiment, the processing precision of the second semantic segmentation processing is higher than that of the first semantic segmentation processing, so that the second semantic segmentation processing can be performed on the basis of the first semantic segmentation result, and the accuracy and the effectiveness of the obtained segmented region can be effectively improved.

In this embodiment, the first semantic segmentation process may be understood as a rough semantic segmentation, and then, on the basis of a result of the rough semantic segmentation, the second semantic segmentation process is performed, where the second semantic segmentation process may be understood as a fine speech segmentation, so that it may be ensured that a second segmentation region obtained by the semantic segmentation is relatively fine and accurate.

S304, according to the at least one second segmentation area, updating the area of interest AOI in the map data corresponding to the first area.

After the at least one second segmentation area is obtained, it can be understood that the area of each geographic entity is segmented in the remote sensing image, and then the AOI area in the map data corresponding to the first area can be updated according to the at least one second segmentation area.

The map data processing method provided by the embodiment of the disclosure comprises the following steps: and acquiring a remote sensing image of the first area. And performing first semantic segmentation processing according to the remote sensing image to obtain at least one first segmentation area. And performing second semantic segmentation processing on the remote sensing image according to the at least one first segmentation region to obtain at least one second segmentation region, wherein the processing precision of the second semantic segmentation processing is higher than that of the first semantic segmentation processing. And updating the area of interest AOI in the map data corresponding to the first area according to the at least one second segmentation area. The method comprises the steps of carrying out semantic segmentation processing with different processing precisions twice on a remote sensing image to obtain at least one second segmentation area, wherein the characteristic of the remote sensing image per se and the semantic segmentation processing of two layers can effectively ensure the accuracy of semantic segmentation, and then updating the AOI area in map data corresponding to the first area according to the second segmentation area obtained by semantic segmentation, so that the efficiency and the accuracy of updating the AOI area can be effectively improved.

To provide a deeper understanding of the reader of the implementation principles of the present disclosure, the embodiment shown in fig. 3 will now be further detailed in conjunction with fig. 4-7 below. Fig. 4 is a second flowchart of a map data processing method provided in the embodiment of the present disclosure, fig. 5 is an implementation schematic diagram of determining a first partition area provided in the embodiment of the present disclosure, fig. 6 is an implementation schematic diagram of determining a mapping area provided in the embodiment of the present disclosure, and fig. 7 is an implementation schematic diagram of determining a second partition area provided in the embodiment of the present disclosure.

As shown in fig. 4, the method includes:

s401, obtaining a remote sensing image of the first area.

The implementation manner of S401 is similar to the implementation manner described in S301, and is not described herein again.

S402, adjusting the image size of the remote sensing image to a preset size to obtain the adjusted remote sensing image.

Based on the above description, it can be determined that the first semantic segmentation process is performed on the remote sensing image in the present disclosure, where the first semantic segmentation process is performed on the entire remote sensing image.

However, since the remote sensing image has a large number of elements, the image size of the remote sensing image is relatively large, and if the image size input to the semantic segmentation network is too large, the network is difficult to learn refined features, and the segmentation accuracy is too low.

Therefore, in this embodiment, the image size of the remote sensing image may be first adjusted to a preset size, where the preset size may be, for example, a size required by the semantic segmentation network, or may also be smaller than the size required by the semantic segmentation network, and this embodiment does not limit an implementation manner of the preset size.

For example, it can be understood with reference to fig. 5, as shown in fig. 5, 501 may be an original remote sensing image, and the remote sensing image 501 is reduced in size to adjust the image size of the remote sensing image 501 to a preset size, so as to obtain an adjusted remote sensing image 502.

S403, performing first semantic segmentation processing on the adjusted remote sensing image to obtain at least one first segmentation region, wherein each first segmentation region corresponds to a respective region type.

After the adjusted remote sensing image is obtained by adjusting the size, the adjusted remote sensing image can be subjected to a first semantic segmentation processing.

In a possible implementation manner, for example, the first semantic segmentation model may be trained in advance in this embodiment, and then, for example, the adjusted remote sensing image may be input into the first semantic segmentation model, and then the first semantic segmentation model may output at least one segmented region. Wherein each first segmentation region corresponds to a respective region class.

The division of the area categories can be selected and set according to actual requirements. For example, the classification of the region categories may depend on data labeling in the model training process, and various types of remote sensing terrain, such as buildings, water areas, roads, vegetation, and the like, may be covered in the data labeling process.

And then training a first semantic segmentation model based on the labeled data. And outputting the segmentation areas of various categories in the remote sensing image by the trained first semantic segmentation model. In practical implementation, the first semantic segmentation model may include different segmentation networks, for example, the segmentation networks may include: deep series, U-Net, laneNet (Lane Detection Networks), segNet, FCN (full volumetric Networks), etc., and this embodiment does not limit the specific implementation manner of the first semantic segmentation model.

For example, as can be understood with reference to fig. 5, it is assumed that, as shown in fig. 5, the first semantic segmentation process is performed on the adjusted remote sensing image 502, and at least 3 segmented regions shown in fig. 5, namely, segmented region 1, segmented region 2, and segmented region 3, are obtained.

It is assumed that the area type corresponding to the divided area 1 is a building, the area type corresponding to the divided area 2 is also a building, and the area type corresponding to the divided area 3 is a road.

In an actual implementation process, the specific division of the first division regions and the region category corresponding to each first division region may be selected and set according to actual requirements.

S404, determining respective circumscribed rectangular areas of the first divided areas in the adjusted remote sensing image.

It will be appreciated that the first semantic segmentation process described above is performed on the resized remote sensing image. However, local features of the remote sensing image are single, and the distinguishing degree of texture features is small. Then the remote sensing image is further resized to result in a small distinction between local feature singleness and texture features.

That is, reducing the size of the remote sensing image causes information loss, and the image also causes pixel errors in the size change process, so that the first semantic segmentation processing on the adjusted remote sensing image cannot ensure the segmentation accuracy.

However, the first semantic segmentation processing can roughly divide the ground elements in the remote sensing image into a plurality of first segmentation areas according to a plurality of pre-classified categories, and then further refine semantic segmentation is performed on each first segmentation area, so that the accuracy of semantic segmentation can be effectively ensured.

Because the first divided region is obtained by dividing the remote sensing image after the size adjustment, and the remote sensing image after the size adjustment loses some details, when the semantic division is further refined for the first divided region, the semantic division is needed to be performed based on the original remote sensing image.

Meanwhile, since the plurality of first segmented regions obtained by semantic segmentation may not be in a regular shape, in order to determine each first semantic segmented region, a corresponding region in the original lumbar rod image, for example, a respective circumscribed rectangular region of each first segmented region may be first determined in the adjusted remote sensing image.

In one possible implementation, for example, the smallest bounding rectangular region of each first partition region may be determined. Alternatively, it is only necessary to ensure that the circumscribed rectangular region can surround the first divided region.

For example, as can be understood with reference to fig. 6, for example, the remote sensing image a after adjustment is determined as the divided

regions

1, 2, and 3. Here, for the divided region 1, a circumscribed rectangular region 601 may be determined, for the divided region 2, a circumscribed rectangular region 602 may be determined, and for the divided region 3, a circumscribed rectangular region 603 may be determined.

S405, determining the mapping area of each circumscribed rectangular area in the original remote sensing image according to the size corresponding relation between the adjusted remote sensing image and the original remote sensing image.

In this embodiment, the adjusted remote sensing image is obtained by resizing based on the original remote sensing image, and therefore there is a size correspondence between the adjusted remote sensing image and the original remote sensing image.

Then, after the circumscribed rectangular regions of the first divided regions are determined in the adjusted remote sensing image, the circumscribed rectangular regions in the adjusted remote sensing image and the corresponding mapping regions in the original remote sensing image can be determined according to the corresponding size relationship between the two images.

For example, as can be understood with reference to fig. 6, for the circumscribed rectangular region 601, a corresponding mapping region 604 may be determined in the original remote sensing image B. And for the circumscribed rectangular region 602, a corresponding mapped region 605 may be determined in the original remote sensing image B. And for the circumscribed rectangular region 603, a corresponding mapping region 606 may be determined in the original remote sensing image B.

And S406, performing second semantic segmentation processing on each mapping region to obtain a second segmentation region corresponding to each mapping region.

After determining the plurality of mapping regions, the second semantic segmentation process may be further performed on each mapping region.

In one possible implementation, for example, the mapping areas may be respectively cropped from the original remote sensing image to obtain a plurality of cropped images. And then, the clipped images can be respectively input into a second semantic segmentation model, and second semantic segmentation processing is respectively carried out, so that second segmentation areas corresponding to the mapping areas are obtained.

Because the mapping area is only a part of area in the remote sensing image, the semantic segmentation processing can be carried out without adjusting the size of the cutting image corresponding to the mapping area, so that the image details in the remote sensing image can be effectively reserved, and the fineness of the semantic segmentation processing is improved.

Meanwhile, the processing precision of the second semantic segmentation processing in this embodiment is higher than that of the first semantic segmentation processing, so that it can be effectively ensured that the segmentation accuracy of the second segmented region obtained by performing the second semantic segmentation processing on the mapping region is relatively high.

In a possible implementation manner, the second semantic segmentation model here is similar to the first semantic segmentation model described above, indicating that the precision requirement on data annotation is higher in the training process of the second semantic segmentation model.

It should be noted here that, in the process of performing the first semantic segmentation process, the reason for reducing the size of the remote sensing image is that the image size of the remote sensing image is relatively large. In the present embodiment, however, semantic segmentation processing is performed on the cropped image corresponding to the mapping region, so that there is no need to adjust the image size.

However, another implementation may be considered here, that is, an original remote sensing image is directly cropped, and then semantic segmentation processing is performed on a plurality of cropped sub-images, so as to ensure that the image size requirement of the semantic segmentation processing can be met on the basis of not reducing the image and reserving the integrity of the image details.

However, the reason why the original remote sensing image is not directly cut is that it is difficult for the model to acquire global information if the image is directly cut and then the cut image is sent to the model for processing. In the remote sensing image, local features of a plurality of objects are similar, and if the size of the directly cut image is too small, the model cannot judge the class of the object in the image.

In this respect, in this embodiment, first semantic segmentation processing is performed on the reduced remote sensing image to obtain a plurality of first segmented regions, where each first segmented region corresponds to a respective region type, so as to ensure that objects in each first segmented region belong to the same type, and meanwhile, in the process of the first semantic segmentation processing, the type of the object is determined first, so as to avoid the problem that the type of the object cannot be determined for the clipped image.

After the first segmentation areas are determined, the corresponding mapping areas in the original remote sensing image are determined for each first segmentation area, and then second semantic segmentation processing is carried out on the mapping areas, so that further refined segmentation is realized, and the segmentation precision is improved.

Therefore, based on the first semantic segmentation processing and the second semantic segmentation processing introduced above, the two layers of processing processes can effectively solve the problem that the remote sensing image is large in size and single in local characteristics, so that fine segmentation is difficult to realize, and the accuracy and fineness of semantic segmentation are effectively improved.

For example, it can be understood in conjunction with fig. 7 that, as shown in fig. 7, assuming that the above-described example is continued, a cropped image 701 is generated for the map area 604, and then the second semantic division process is performed on the cropped image 701.

It can be determined based on the above description that the area category corresponding to the mapping area is 604 a building, but it can be determined from the figure that a plurality of buildings are actually included in the mapping area, because the first semantic segmentation process is performed on the reduced remote sensing image, the connected buildings cannot be distinguished, and are segmented into the same area.

Since the second semantic division processing is performed on the clipped region in the original remote sensing image and the second semantic division processing is more accurate, referring to fig. 7, performing the second semantic division processing on 701 can further divide a plurality of buildings included in the region, and obtain a second divided region 702, a second divided region 703, a second divided region 704, a second divided region 705, and a second divided region 706 shown in fig. 7.

After each mapping region is processed to obtain a corresponding second segmentation region, a plurality of second segmentation regions corresponding to the remote sensing image are actually obtained. In a possible implementation manner, for example, the segmentation results of the multiple mapping regions may be further spliced to obtain a final semantic segmentation result corresponding to the remote sensing image, and the final semantic segmentation result includes the multiple second segmentation regions described above.

And S407, determining coordinate information of each second division area in a second coordinate system according to the conversion relation between the first coordinate system corresponding to the remote sensing image and the second coordinate system corresponding to the map data.

After semantic segmentation is performed on the remote sensing image to obtain a plurality of second segmentation areas, the AOI area in the map data can be updated according to the plurality of second segmentation areas.

However, since the remote sensing image and the map data are data located in different coordinate systems, comparison of regions or pixels cannot be directly performed without processing.

Therefore, in this embodiment, the coordinate information of each second divided area in the second coordinate system may be determined first according to the coordinate system conversion relationship between the first coordinate system corresponding to the remote sensing image and the second coordinate system corresponding to the map data. That is, a plurality of second divided areas in the remote sensing image are converted into a second coordinate system corresponding to the map data, so as to facilitate subsequent comparison processing.

In a possible implementation manner, for example, the coordinate system of each pixel in the second divided region is converted to obtain coordinate information of the second divided region in the second coordinate system, and the coordinate information includes coordinate values of each pixel in the second divided region in the second coordinate system. Or, the coordinate system may be converted for each pixel point on the boundary of the second divided region, so as to obtain the coordinate information of the second divided region in the second coordinate system, where the coordinate information includes the coordinate value of each pixel point on the boundary of the second divided region in the second coordinate system.

And S408, determining first contour information according to the coordinate information of each second division area, wherein the first contour information comprises the area contour of each second division area.

After the coordinate information of the second divided region is determined, the contour information of the second divided region may be determined according to the coordinate information of the second divided region, for example, the first contour information of the second divided region may be determined according to pixel points on the boundary of the second divided region where coordinate values are connected, and the first contour information may include the region contour of each second divided region.

Or, when determining the contour information, for example, the contour information may be determined through techniques such as edge extraction and filtering, and the specific implementation manner of determining the contour information of a certain area is not limited in this embodiment, and may be selected and set according to actual requirements.

In a possible implementation manner, the first contour information in this embodiment may be information in the form of an image, that is, a region contour of the second divided region is labeled in the image. Alternatively, the first contour information in this embodiment may also be information in a data format, for example, coordinate values of each pixel point may be sequentially stored according to connectivity of the pixel points on the boundary, so as to label the area contour of the second divided area.

And S409, acquiring second contour information, wherein the second contour information comprises the area contour of each AOI area in the map data.

In order to compare the AOI area in the map data with the second divided area, the embodiment also needs to acquire the area outline of each AOI area existing in the map data, that is, the second outline information. The implementation manner of the second profile information is similar to the implementation manner of the first profile information described above, and is not described herein again.

S410, performing a difference processing according to the first contour information and the second contour information to obtain a difference result of the first contour information relative to the second contour information, where the difference result includes at least one of the following: new region profile, missing region profile, updated region profile.

After the first contour information and the second contour information are determined, wherein the first contour information represents a plurality of semantic segmentation areas obtained by semantic segmentation according to the remote sensing image, and the second contour information represents a plurality of AOI areas existing in the current map data, differential processing can be carried out according to the first contour information and the second contour information, so that a differential result of the first contour information relative to the second contour information is obtained. The difference result here can also be understood as the difference where the first profile information exists with respect to the second profile information.

In a possible implementation manner, the difference result in this embodiment may include at least one of the following: new region profile, missing region profile, updated region profile.

The new region profile is understood to mean the region profile present in the second segmented region, but not in the AOI region.

And, missing region contours can be understood as region contours that are not present in the second segmented region, but are present in the AOI region.

And updating the region profile may be understood as the region profile is present in both the second segmentation region and the AOI region, but the shape of the region profile is changed.

And S411, updating the AOI area in the map data according to the difference result.

After the difference result described above is obtained, it can be determined what the newly divided second divided region has changed from the currently existing AOI region, so that the existing AOI region in the map data can be updated according to the difference result.

The three difference results described above are described below with reference to specific cases:

the new region profile corresponds to a region profile of a geographic entity newly added in an actual scene, so that a second divided region corresponding to the new region profile can be added to the map data as a newly added AOI region.

And for the missing area outline, the area outline actually corresponding to the geographic entity removed in the actual scene is the area outline, so for the missing area outline, the AOI area corresponding to the missing area outline can be deleted in the map data.

And for the updated area profile, what actually corresponds to the area profile of the geographic entity whose floor area has changed in the actual scene, for example, the park is extended or reduced, and so on, so for the updated area profile, the second partition area corresponding to the updated area profile may be adopted to replace the AOI area corresponding to the updated area profile in the map data.

In an alternative implementation manner, before the AOI area in the map data is updated according to the difference result, for example, the difference result may be sent to the target terminal device, where the target terminal device may be understood as a terminal device corresponding to a worker, that is, the worker further verifies the difference result.

When the staff determines that the difference result is not problematic, the confirmation information corresponding to the difference result may be sent to the operating device in this embodiment through the target terminal device, and after receiving the confirmation information, the AOI area in the map data is updated according to the difference result.

And it should be further noted that the AOI information in the electronic map includes four items of information, namely, name, address, category, and longitude and latitude.

In this embodiment, when the AOI in the map data is updated, the important point is to update the AOI area, or in an optional implementation manner, the address and/or the latitude and longitude of the AOI may also be updated according to the geographic location information included in the remote sensing image, but more detailed information, for example, the name of the AOI, needs to be supplemented by a worker for improvement.

According to the map data processing method provided by the embodiment of the disclosure, the image size of the remote sensing image is firstly adjusted, then the first semantic segmentation processing is firstly performed on the basis of the adjusted remote sensing image, so as to obtain the first segmentation areas corresponding to respective area categories, thereby realizing the first semantic segmentation on the basis of the complete remote sensing image, and effectively determining the categories of the respective segmentation areas. And then, aiming at each first segmentation area, determining a corresponding mapping area in the original remote sensing image, and then aiming at each mapping area, performing second semantic segmentation processing with higher processing precision, wherein the mapping area is only a part of area in the remote sensing image, so that the semantic segmentation can be directly performed without performing size adjustment on the mapping area, the size of the mapping area is not reduced, and the processing precision of the second semantic segmentation processing is higher than that of the first semantic segmentation processing, so that the refined segmentation of each geographic element in the remote sensing image can be effectively realized, and the accuracy and the effectiveness of a semantic segmentation result are improved. After the second semantic segmentation result of the remote sensing image is determined, the coordinate information of each second segmentation area under the second coordinate system can be determined according to the conversion relation between the coordinate system of the remote sensing image and the coordinate system of the map data, so that the second segmentation area and the AOI area are ensured to be located under the same coordinate system, and subsequent area differential processing is facilitated. Then, the difference processing can be performed according to the first profile information corresponding to the second divided area and the second profile information corresponding to the AOI area, and then the AOI area in the map data is updated according to the difference result. Meanwhile, the process is automatically operated, so that the processing efficiency of updating the AOI information can be effectively improved.

Fig. 8 is a schematic structural diagram of a map data processing apparatus according to an embodiment of the present disclosure. As shown in fig. 8, the map data processing apparatus 800 of the present embodiment may include: an acquisition module 801, a first processing module 802, a second processing module 803, and an update module 804.

An obtaining module 801, configured to obtain a remote sensing image of a first area;

a first processing module 802, configured to perform a first semantic segmentation processing according to the remote sensing image to obtain at least one first segmentation region;

a second processing module 803, configured to perform second semantic segmentation processing on the remote sensing image according to the at least one first segmentation region, so as to obtain at least one second segmentation region, where processing accuracy of the second semantic segmentation processing is higher than that of the first semantic segmentation processing;

an updating module 804, configured to update, according to the at least one second divided area, an AOI area of the interest plane in the map data corresponding to the first area.

In a possible implementation manner, the first processing module 802 is specifically configured to:

adjusting the image size of the remote sensing image to a preset size to obtain an adjusted remote sensing image;

and carrying out first semantic segmentation processing on the adjusted remote sensing image to obtain at least one first segmentation region, wherein each first segmentation region corresponds to a respective region type.

In a possible implementation manner, the second processing module 803 is specifically configured to:

determining respective circumscribed rectangular areas of the first segmentation areas in the adjusted remote sensing image;

determining mapping areas corresponding to the circumscribed rectangular areas in the original remote sensing image according to the corresponding size relation between the adjusted remote sensing image and the original remote sensing image;

and respectively carrying out second semantic segmentation processing on each mapping region to obtain a second segmentation region corresponding to each mapping region.

In a possible implementation manner, the updating module 804 is specifically configured to:

determining coordinate information of each second division area in a second coordinate system according to a conversion relation between the first coordinate system corresponding to the remote sensing image and the second coordinate system corresponding to the map data;

determining first contour information according to the coordinate information of each second division area, wherein the first contour information comprises the area contour of each second division area;

acquiring second contour information, wherein the second contour information comprises area contours of all AOI areas existing in the map data;

and updating the AOI area in the map data according to the first contour information and the second contour information.

performing difference processing according to the first contour information and the second contour information to obtain a difference result of the first contour information relative to the second contour information, wherein the difference result includes at least one of the following: newly adding a region profile, missing a region profile and updating a region profile;

and updating the AOI area in the map data according to the difference result.

In a possible implementation manner, the update module 804 is specifically configured to:

according to the new area contour in the difference result, taking a second divided area corresponding to the new area contour as a new AOI area, and adding the new AOI area into the map data; and (c) a second step of,

deleting the AOI area corresponding to the missing contour area in the map data according to the contour of the missing area in the difference result; and the number of the first and second groups,

and replacing the AOI area corresponding to the updated area outline in the map data by adopting a second segmentation area corresponding to the updated area outline according to the updated area outline in the difference result.

In a possible implementation manner, the apparatus further includes: a transmission module 805;

the transmission module 805 is configured to send the difference result to a target terminal device before updating the AOI area in the map data according to the difference result;

and receiving confirmation information corresponding to the difference result sent by the target terminal equipment, wherein the confirmation information is used for indicating that the AOI area in the map data is updated according to the difference result.

The invention provides a map data processing method and device, which are applied to the field of car networking and intelligent cabins in the field of data processing so as to achieve the purpose of improving the efficiency and accuracy of updating AOI in map data.

It should be noted that the head model in this embodiment is not a head model for a specific user, and cannot reflect personal information of a specific user. It should be noted that the two-dimensional face image in the present embodiment is derived from a public data set.

In the technical scheme of the disclosure, the collection, storage, use, processing, transmission, provision, disclosure and other processing of the personal information of the related user are all in accordance with the regulations of related laws and regulations and do not violate the good customs of the public order.

The present disclosure also provides an electronic device, a readable storage medium, and a computer program product according to embodiments of the present disclosure.

According to an embodiment of the present disclosure, the present disclosure also provides a computer program product comprising: a computer program, stored in a readable storage medium, from which at least one processor of the electronic device can read the computer program, and the execution of the computer program by the at least one processor causes the electronic device to perform the solutions provided by any of the above embodiments.

FIG. 9 illustrates a schematic block diagram of an example electronic device 900 that can be used to implement embodiments of the present disclosure. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. Electronic devices may also represent various forms of mobile devices, such as personal digital processors, cellular telephones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the disclosure described and/or claimed herein.

As shown in fig. 9, the device 900 comprises a computing unit 901 which may perform various appropriate actions and processes in accordance with a computer program stored in a read-only memory (ROM) 902 or a computer program loaded from a storage unit 908 into a Random Access Memory (RAM) 903. In the RAM 903, various programs and data required for the operation of the device 900 can also be stored. The calculation unit 901, ROM 902, and RAM 903 are connected to each other via a bus 904. An input/output (I/O) interface 905 is also connected to bus 904.

A number of components in the device 900 are connected to the I/O interface 905, including: an input unit 906 such as a keyboard, a mouse, and the like; an output unit 907 such as various types of displays, speakers, and the like; a storage unit 908 such as a magnetic disk, optical disk, or the like; and a communication unit 909 such as a network card, a modem, a wireless communication transceiver, and the like. The communication unit 909 allows the device 900 to exchange information/data with other devices through a computer network such as the internet and/or various telecommunication networks.

The computing unit 901 may be a variety of general and/or special purpose processing components with processing and computing capabilities. Some examples of the computing unit 901 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various dedicated Artificial Intelligence (AI) computing chips, computing units running various machine learning model algorithms, a Digital Signal Processor (DSP), and any suitable processor, controller, microcontroller, and the like. The calculation unit 901 performs the respective methods and processes described above, such as the map number processing method. For example, in some embodiments, the map number processing method may be implemented as a computer software program tangibly embodied in a machine-readable medium, such as storage unit 908. In some embodiments, part or all of the computer program may be loaded and/or installed onto device 900 via ROM 902 and/or communications unit 909. When the computer program is loaded into the RAM 903 and executed by the computing unit 901, one or more steps of the map number processing method described above may be performed. Alternatively, in other embodiments, the computing unit 901 may be configured to perform the map number processing method by any other suitable means (e.g., by means of firmware).

Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuitry, field Programmable Gate Arrays (FPGAs), application Specific Integrated Circuits (ASICs), application Specific Standard Products (ASSPs), system on a chip (SOCs), complex Programmable Logic Devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs that are executable and/or interpretable on a programmable system including at least one programmable processor, which may be special or general purpose, receiving data and instructions from, and transmitting data and instructions to, a storage system, at least one input device, and at least one output device.

Program code for implementing the methods of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the program codes, when executed by the processor or controller, cause the functions/acts specified in the flowchart and/or block diagram block or blocks to be performed. The program code may execute entirely on the machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.

In the context of this disclosure, a machine-readable medium may be a tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which a user can provide input to the computer. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic, speech, or tactile input.

The systems and techniques described here can be implemented in a computing system that includes a back-end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back-end, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), wide Area Networks (WANs), and the Internet.

The computer system may include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The Server can be a cloud Server, also called a cloud computing Server or a cloud host, and is a host product in a cloud computing service system, so as to solve the defects of high management difficulty and weak service expansibility in the traditional physical host and VPS service ("Virtual Private Server", or simply "VPS"). The server may also be a server of a distributed system, or a server incorporating a blockchain.

It should be understood that various forms of the flows shown above may be used, with steps reordered, added, or deleted. For example, the steps described in the present disclosure may be executed in parallel, sequentially, or in different orders, as long as the desired results of the technical solutions disclosed in the present disclosure can be achieved, and the present disclosure is not limited herein.

The above detailed description should not be construed as limiting the scope of the disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may be made, depending on design requirements and other factors. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present disclosure should be included in the protection scope of the present disclosure.

Claims

1. A map data processing method, comprising:

acquiring a remote sensing image of a first area;

performing first semantic segmentation processing according to the remote sensing image to obtain at least one first segmentation area;

and updating the area of interest (AOI) in the map data corresponding to the first area according to the at least one second segmentation area.

2. The method of claim 1, wherein said performing a first semantic segmentation process from the remote sensing image to obtain at least one first segmented region comprises:

3. The method of claim 2, wherein said performing a second semantic segmentation process on the remote sensing image according to the at least one first segmented region to obtain at least one second segmented region comprises:

determining respective circumscribed rectangular areas of the first divided areas in the adjusted remote sensing image;

4. The method according to any one of claims 1 to 3, wherein the updating, according to the at least one second divided area, the area of interest AOI area in the map data corresponding to the first area includes:

5. The method of claim 4, wherein the updating the AOI area in the map data according to the first and second profile information comprises:

and updating the AOI area in the map data according to the difference result.

6. The method of claim 5, wherein the updating the AOI region in the map data according to the difference result comprises:

deleting the AOI area corresponding to the missing outline area in the map data according to the missing area outline in the difference result; and the number of the first and second groups,

7. The method according to claim 5 or 6, before updating the AOI area in the map data according to the difference result, the method further comprising:

sending the difference result to target terminal equipment;

8. A map data processing apparatus comprising:

the second processing module is used for carrying out second semantic segmentation processing on the remote sensing image according to the at least one first segmentation region to obtain at least one second segmentation region, wherein the processing precision of the second semantic segmentation processing is higher than that of the first semantic segmentation processing;

9. The apparatus of claim 8, wherein the first processing module is specifically configured to:

and carrying out first semantic segmentation processing on the adjusted remote sensing image to obtain at least one first segmentation area, wherein each first segmentation area corresponds to a respective area category.

10. The apparatus according to claim 9, wherein the second processing module is specifically configured to:

11. The apparatus according to any one of claims 8-10, wherein the update module is specifically configured to:

12. The apparatus of claim 11, wherein the update module is specifically configured to:

and updating the AOI area in the map data according to the difference result.

13. The apparatus of claim 12, wherein the update module is specifically configured to:

aiming at the newly added area outline in the difference result, taking a second divided area corresponding to the newly added area outline as a newly added AOI area, and adding the newly added area outline into the map data; and the number of the first and second groups,

deleting the AOI area corresponding to the missing contour area in the map data according to the missing contour in the difference result; and the number of the first and second groups,

14. The apparatus of claim 12 or 13, further comprising: a transmission module;

the transmission module is used for sending the difference result to target terminal equipment before the AOI area in the map data is updated according to the difference result;

15. An electronic device, comprising:

at least one processor; and

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-7.

16. A non-transitory computer readable storage medium having stored thereon computer instructions for causing the computer to perform the method of any one of claims 1-7.

17. A computer program product comprising a computer program which, when executed by a processor, carries out the steps of the method of any one of claims 1 to 7.