CN117292394B

CN117292394B - Map auditing method and device

Info

Publication number: CN117292394B
Application number: CN202311266912.9A
Authority: CN
Inventors: 左栋; 梁宇; 狄琳; 张雨心; 王紫玉; 吴彬; 杨建忠; 张通滨; 侯燕
Original assignee: Map Technology Examination Center Of Ministry Of Natural Resources
Current assignee: Map Technology Examination Center Of Ministry Of Natural Resources
Priority date: 2023-09-27
Filing date: 2023-09-27
Publication date: 2024-04-30
Anticipated expiration: 2043-09-27
Also published as: CN117292394A

Abstract

The disclosure provides a map auditing method and device, relates to the technical field of artificial intelligence, and particularly relates to the technical fields of computer vision, image processing, deep learning and the like. One embodiment of the method comprises the following steps: acquiring a map to be checked; selecting a true value map corresponding to the map to be checked; constructing an image feature set of the map to be checked and an image feature set of the truth value map based on a multi-scale line convolution and a feature extraction algorithm of the surface convolution; and comparing the image feature set of the map to be checked with the image feature set of the truth value map to obtain a checking result of the map to be checked. According to the embodiment, the abnormal points of the map to be checked can be obtained by comparing the image feature set of the map to be checked with the image feature set of the truth value map, so that automatic checking is realized.

Description

Map auditing method and device

Technical Field

The present disclosure relates to the technical field of artificial intelligence, and in particular to the technical fields of computer vision, image processing, deep learning, and the like.

Background

A map is a graphic or image that represents several phenomena of the earth (or other star) on a plane or sphere, optionally in two or more dimensions and means, according to a certain law. The method has strict mathematical foundation, symbol system and text annotation, and can scientifically reflect the distribution characteristics of natural and socioeconomic phenomena and the interrelationship thereof by using a map summarizing principle. The map is checked, so that the occurrence of the problem map can be effectively reduced.

Currently, map auditing is mainly performed by manual map auditing. Aiming at the map to be checked provided by the checking unit, an operator manually searches a truth value map on a checking platform; the map to be checked is compared with the true value map by naked eyes, and the map position with problems is found out; and manually returning the map suspected to be problematic after the verification to a verification unit for correction, and re-verifying after the correction.

Disclosure of Invention

The embodiment of the disclosure provides a map auditing method, device, equipment, storage medium and program product.

In a first aspect, an embodiment of the present disclosure provides a map auditing method, including: acquiring a map to be checked; selecting a true value map corresponding to the map to be checked; constructing an image feature set of the map to be checked and an image feature set of the truth value map based on a multi-scale line convolution and a feature extraction algorithm of the surface convolution; and comparing the image feature set of the map to be checked with the image feature set of the truth value map to obtain a checking result of the map to be checked.

In a second aspect, an embodiment of the present disclosure provides a map auditing apparatus, including: the acquisition module is configured to acquire the map to be checked; the selection module is configured to select a true value map corresponding to the map to be checked; the construction module is configured to construct an image feature set of the map to be checked and an image feature set of the truth value map based on a multi-scale line convolution and a feature extraction algorithm of the surface convolution; the comparison module is configured to compare the image feature set of the map to be checked with the image feature set of the truth value map to obtain a check result of the map to be checked.

In a third aspect, an embodiment of the present disclosure proposes an electronic device, including: at least one processor; and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method as described in the first aspect.

In a fourth aspect, embodiments of the present disclosure provide a non-transitory computer-readable storage medium storing computer instructions for causing a computer to perform a method as described in the first aspect.

In a fifth aspect, embodiments of the present disclosure propose a computer program product comprising a computer program which, when executed by a processor, implements a method as described in the first aspect.

According to the map auditing method provided by the embodiment of the disclosure, the abnormal points of the map to be audited can be obtained by comparing the image feature set of the map to be audited with the image feature set of the truth map, so that automatic auditing is realized. And the efficiency and accuracy of map auditing are greatly improved.

Nor is it intended to limit the scope of the present disclosure to the critical or important features of the embodiments of the present disclosure. Other features of the present disclosure will become apparent from the following specification.

Drawings

Other features, objects and advantages of the present disclosure will become more apparent upon reading of the detailed description of non-limiting embodiments made with reference to the following drawings. The drawings are for a better understanding of the present solution and are not to be construed as limiting the present disclosure. Wherein:

FIG. 1 is an exemplary system architecture diagram in which the present application may be applied;

FIG. 2 is a flow chart of one embodiment of a map auditing method according to the present disclosure;

FIG. 3 is a flow chart of yet another embodiment of a map auditing method according to the present disclosure;

FIG. 4 is a schematic diagram of a conventional one-dimensional convolution;

FIG. 5 is a schematic diagram of a multi-scale line convolution;

FIG. 6 is a schematic illustration of a face convolution;

FIG. 7 is yet another schematic of a face convolution;

FIG. 8 is a scene diagram of a map auditing method in which embodiments of the present disclosure may be implemented;

FIG. 9 is a schematic structural view of one embodiment of a map auditing apparatus according to the present disclosure;

fig. 10 is a block diagram of an electronic device for implementing a map auditing method of an embodiment of the present disclosure.

Detailed Description

Exemplary embodiments of the present disclosure are described below in conjunction with the accompanying drawings, which include various details of the embodiments of the present disclosure to facilitate understanding, and should be considered as merely exemplary. Accordingly, one 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.

It should be noted that, without conflict, the embodiments of the present disclosure and features of the embodiments may be combined with each other. The present disclosure will be described in detail below with reference to the accompanying drawings in conjunction with embodiments.

FIG. 1 illustrates an exemplary system architecture 100 in which embodiments of a map auditing method or map auditing apparatus of the present application may be applied.

As shown in fig. 1, system architecture 100 may include a terminal device 101, a network 102, and a server 103. Network 102 is the medium used to provide communication links between terminal device 101 and server 103. Network 102 may include various connection types such as wired, wireless communication links, or fiber optic cables, among others.

The review entity may interact with the server 103 via the network 102 using the terminal device 101 to receive or send messages or the like. For example, the review sending unit may provide a review sending map to the server 103 using the terminal device 101, and receive a map review result returned by the server 103.

The terminal device 101 may be hardware or software. When the terminal device 101 is hardware, it may be a variety of electronic devices including, but not limited to, smartphones, tablets, laptop and desktop computers, and the like. When the terminal apparatus 101 is software, it may be installed in the above-described electronic apparatus. Which may be implemented as a plurality of software or software modules, or as a single software or software module. The present invention is not particularly limited herein.

The server 103 may be a server providing various services, such as a map auditing server. The map review server may analyze and process the received data such as the review map and feed back the processing result (e.g., the map review result) to the terminal device 101.

The server 103 may be hardware or software. When the server 103 is hardware, it may be implemented as a distributed server cluster composed of a plurality of servers, or may be implemented as a single server. When the server 103 is software, it may be implemented as a plurality of software or software modules (for example, to provide distributed services), or may be implemented as a single software or software module. The present invention is not particularly limited herein.

It should be noted that, the map auditing method provided in the embodiment of the present application is generally executed by the server 103, and accordingly, the map auditing device is generally disposed in the server 103.

It should be understood that the number of terminal devices, networks and servers in fig. 1 is merely illustrative. There may be any number of terminal devices, networks, and servers, as desired for implementation.

FIG. 2 illustrates a flow 200 of one embodiment of a map auditing method according to the present disclosure. The map auditing method comprises the following steps:

step 201, obtaining a map to be checked.

In this embodiment, based on the received review map, a review map may be obtained.

Wherein, the censoring unit can provide censoring map. Specifically, based on the review machine provided by the review unit, the review map can be acquired. The review map can be directly used as a review map, or can be further processed to obtain the review map. For example, the review sending map is automatically intercepted through an automatic screenshot tool and map element coordinate points, and a to-be-reviewed map is obtained. In this way, the review map is generally cut into individual area maps, and each area map is reviewed.

And 202, selecting a true value map corresponding to the map to be checked.

In this embodiment, based on the map to be checked, a true value map corresponding thereto may be selected.

The truth map may be a standard map in which no abnormality exists. The truth map corresponding to the review map may be a truth map consistent with the review map style. The style may include, but is not limited to, at least one of: typesetting, color, hue, and region, and so forth.

In some embodiments, a truth map consistent with the style of the map under review may be selected from a truth gallery. The truth map library may store a plurality of truth maps. The truth value maps can be derived from maps passing history examination, so that the truth value map library is expanded, and data is accumulated for subsequent automatic examination work.

And 203, constructing an image feature set of the map to be checked and an image feature set of the truth value map based on a multi-scale line convolution and a feature extraction algorithm of the surface convolution.

In this embodiment, based on the map to be checked, a feature extraction algorithm based on multi-scale line convolution and surface convolution may be used to construct an image feature set of the map to be checked. Similarly, based on the truth value map, a feature extraction algorithm based on multi-scale line convolution and surface convolution can be used for constructing an image feature set of the truth value map.

The feature extraction algorithm of the multi-scale line convolution can deflect the convolution kernel with at least one size by utilizing at least one deflection angle, and then convolve the image by utilizing the deflected convolution kernel. The feature extraction algorithm of the multi-scale surface convolution can utilize at least one span to migrate the features, and then the original features at the same position and the migrated features are fused. The image feature set may include a number of features of the map. These features may be extracted by convolving the map, including but not limited to: the national boundary line, the administrative region boundary line of each level, the important water river, the shape and the ground color of each region of the map, the text label of the administrative region of each level, the ground color of each administrative region and the like.

And 204, comparing the image feature set of the map to be checked with the image feature set of the truth value map to obtain a checking result of the map to be checked.

In this embodiment, the auditing result of the map to be audited may be obtained by comparing the image feature set of the map to be audited with the image feature set of the truth value map.

For one feature in the image feature set of the map to be checked, the corresponding feature can be found out from the image feature set of the truth value map, and the two features are matched. If the two features are matched, the corresponding coordinate points are not abnormal; if the two features are not matched, the corresponding coordinate points are abnormal.

The auditing result can be used for representing whether the map to be audited passes or not. Typically, no map audit of outliers is passed. And for the map which passes the examination, the map can be added into a truth map library, so that the truth map library is expanded, and data is accumulated for subsequent automatic examination work. Map audits with outliers are not passed. And outputting the map which fails to pass the audit to the audit unit for repairing and correcting. After repair and correction by the delivery unit, the map auditing method provided by the embodiment of the disclosure can be executed again to carry out next round of auditing.

In some embodiments, the outlier region of the map to be reviewed may be obtained by clustering outliers. And marking the abnormal region of the map to be audited, and outputting the marked abnormal region to an audit sending unit. Therefore, the sending and checking unit can repair and correct the map in a targeted manner, and the repair and correction efficiency of the map is improved.

It should be noted that, the execution body of the map auditing method provided in the embodiment of the present disclosure may be the server 103 in the system architecture 100 shown in fig. 1, and specifically, refer to the embodiment shown in fig. 1, which is not described herein again.

According to the map auditing method provided by the embodiment of the disclosure, the abnormal points of the map to be audited can be obtained by comparing the image feature set of the map to be audited with the image feature set of the truth map, so that automatic auditing is realized. The map auditing cost is reduced, and the map auditing efficiency and accuracy are improved.

With continued reference to fig. 3, a flow 300 of yet another embodiment of a map auditing method according to the present disclosure is shown. The map auditing method comprises the following steps:

Step 301, obtaining a map to be checked.

And 302, selecting a true value map corresponding to the map to be checked.

In this embodiment, the specific operations of steps 301 to 302 are described in detail in steps 201 to 202 in the embodiment shown in fig. 2, and are not described herein.

And 303, respectively carrying out multi-scale line convolution on the map to be checked and the truth value map to obtain a first image feature subset of the map to be checked and a first image feature subset of the truth value map.

In this embodiment, the multi-scale line convolution is performed on the map to be checked and the truth value map respectively, so that a first image feature subset of the map to be checked and a first image feature subset of the truth value map can be obtained. The line convolution can extract image features of multiple dimensions such as lines, planes and the like.

Fig. 4 shows a schematic diagram of a conventional one-dimensional convolution. Taking the example of a convolution kernel size equal to 4, there are two forms of 1×4 and 4×1.

Fig. 5 shows a schematic diagram of a multi-scale line convolution. Multi-scale line convolution introduces a parameter of deflection angle in addition to defining the convolution kernel size. A polar coordinate system is established by using one pixel point, and fig. 5 shows two different deflection angles θ ₁ and θ ₂ under different polar coordinate systems, and the gray pixel point covered by the two different deflection angles θ ₁ and θ ₂ is the shape of the line convolution. And when the deflection angles are 0, 90, 180 and 270, the traditional one-dimensional convolution kernel can be obtained. In the case of a fixed size convolution kernel, the entire circular area can be covered by defining the angle of deflection of the line convolution kernel.

In some embodiments, the multi-scale line convolution is performed on the map to be audited and the truth map based on at least one convolution kernel size and at least one deflection angle, respectively, to obtain a first image feature subset of the map to be audited and a first image feature subset of the truth map. The convolution kernel size and the deflection angle can be preset according to the convolution times of the multi-scale line convolution. A combination of a convolution kernel size and a deflection angle may be used to perform a one-scale line convolution. By setting at least one convolution kernel size and at least one deflection angle, a multi-scale line convolution can be performed, thereby extracting more line elements and face elements.

In general, the number of convolution kernel sizes and the number of deflection angles can be flexibly adjusted according to requirements. If the map content is simpler, the map auditing can be completed by only extracting a small number of image features, and only a small number of convolution kernel sizes and deflection angles are required to be set. For example, two convolution kernel sizes and three deflection angles are set, and the convolution kernel sizes and the deflection angles are combined two by two to perform line convolution of six scales. If the map content is complex, a large number of image features need to be extracted to complete the map audit, and a large number of convolution kernel sizes and deflection angles need to be set. For example, three convolution kernel sizes and twelve deflection angles are set, and the convolution kernel sizes and the deflection angles are combined two by two to perform line convolution of thirty-six scales.

In the map verification, line elements and face elements are important bases for judging whether a map is compliant, such as national boundaries, administrative region boundaries of each level, important water-based rivers, shape and ground color of each region of the map, and the like. The multi-scale line convolution can perform linear convolution calculation from various angles, can better adapt to complex features such as shapes, styles, trend, width and the like of various map line elements, and is focused on extraction of map linear image features.

And step 304, respectively carrying out multi-scale convolution on the map to be checked and the truth value map to obtain a second image feature subset of the map to be checked and a second image feature subset of the truth value map.

In this embodiment, the multi-scale plane convolution is performed on the map to be checked and the truth value map respectively, so that a second image feature subset of the map to be checked and a second image feature subset of the truth value map can be obtained. Wherein the face convolution can extract image features of the face dimension.

Based on the characteristic diagram of CxH x W (CHANNEL HIGH WIDTH, channel height width) obtained by the traditional two-dimensional convolution calculation, the face convolution adds a characteristic fusion process.

In some embodiments, slicing is performed on the feature map of the map to be checked and the feature map of the truth value map according to the height direction and the width direction, so that a feature slice map of the map to be checked and a feature slice map of the truth value map can be obtained; and respectively migrating and fusing the features in the feature slice diagram of the map to be checked and the features in the feature slice diagram of the truth value map to a preset direction according to at least one migration span, so as to obtain a second image feature subset of the map to be checked and a second image feature subset of the truth value map. The migration span may be preset according to the convolution times of the multi-scale surface convolution. A migration span may be one-scale surface convolution. By setting at least one migration span, a multi-scale surface convolution can be performed, thereby extracting more surface elements. By means of feature fusion, face convolution can fuse the features of all areas together more quickly. By constructing a more complex mapping relationship, the characteristics of the surface elements such as the text labels of all levels of administrative division and the ground colors of all administrative areas in map auditing can be better extracted.

In general, the number of migration spans can be flexibly adjusted according to requirements. If the map content is simpler, the map auditing can be completed by only extracting a small number of image features, and only a small number of migration spans are required to be set. If the map content is complex, a large number of image features need to be extracted to complete the map audit, and a large number of migration spans need to be set.

Fig. 6 shows a schematic diagram of a face convolution. As shown in fig. 6, the feature map is sliced in the H direction (i.e., the height direction), and then the sliced features are sequentially migrated and fused from left to right according to a migration span of 2. Specifically, features of the first column are migrated from left to right to the third column according to a migration span of 2, and original features of the third column are fused with features migrated to the third column to replace the features of the third column. And migrating the features of the second column from left to right to the fourth column according to the migration span of 2, and fusing the original features of the fourth column with the features migrated to the fourth column to replace the features of the fourth column. And so on until all the features are migrated and fused.

Fig. 7 shows a further schematic of a face convolution. As shown in fig. 7, the feature map is sliced in the W direction (i.e., the width direction), and then the sliced features are sequentially migrated and fused from bottom to top in accordance with a migration span of 2. Specifically, features of the first column are migrated from bottom to top to the third column according to a migration span of 2, and original features of the third column are fused with features migrated to the third column to replace the features of the third column. And migrating the features of the second column from bottom to top to the fourth column according to the migration span of 2, and fusing the original features of the fourth column with the features migrated to the fourth column to replace the features of the fourth column. And so on until all the features are migrated and fused.

Step 305, merging the first image feature subset and the second image feature subset of the map to be checked to obtain an image feature set of the map to be checked, and merging the first image feature subset and the second image feature subset of the truth map to obtain an image feature set of the truth map.

In this embodiment, the first image feature subset and the second image feature subset of the map to be checked are combined, so that the image feature set of the map to be checked can be obtained. Similarly, the first image feature subset and the second image feature subset of the truth map are combined, and the image feature set of the truth map can be obtained.

And 306, comparing the image feature set of the map to be checked with the image feature set of the truth value map by using the self-adaptive feature matching network to obtain an abnormal point detection result of the map to be checked.

In this embodiment, the self-adaptive feature matching network is used to compare the image feature set of the map to be checked with the image feature set of the truth map to obtain the abnormal point detection result of the map to be checked. Specifically, the image feature set of the map to be checked and the image feature set of the truth value map are respectively input into the self-adaptive feature matching network, so that an abnormal point detection result of the map to be checked can be obtained.

The self-adaptive feature matching network can match the image feature set of the map to be checked with the image feature set of the truth value map based on the self-adaptive feature matching algorithm, so that whether abnormal points exist in the map to be checked is detected.

In some embodiments, the adaptive feature matching network comprises at least one of a CNN (Convolutional neural network ) and a transducer (transducer model), with which features in the image feature set of the map under review and features in the image feature set of the truth map are converted into feature vectors, respectively; and matching the feature vector of the map to be checked with the corresponding feature vector of the truth value map to obtain an abnormal point detection result of the map to be checked. The CNN or the transducer can abstract the multidimensional features into one-dimensional feature vectors, so that subsequent matching calculation is facilitated.

Wherein the matching of the feature vectors may be performed from at least one of the following dimensions: firstly, directly matching, and calculating the similarity between the feature vector of the map to be checked and the corresponding feature vector of the true value map; performing coordinate reduction on the feature vectors with the similarity lower than a preset threshold value of the similarity to obtain abnormal points of the map to be checked; secondly, indirectly matching, calculating the difference value between the feature vector of the map to be checked and the feature vector of the surrounding area, and the difference value between the corresponding feature vector of the truth value map and the feature vector of the surrounding area, and comparing the difference value between the two difference values; and carrying out coordinate reduction on the feature vectors with the difference larger than the difference preset threshold value to obtain abnormal points of the map to be checked.

In practical application, the feature vector matching in the two modes is usually carried out, so that more abnormal points can be matched, and the recall rate of the problem map is improved.

Step 307, determining whether the map to be checked has an outlier.

In the present embodiment, based on the abnormal point detection result, it may be determined whether or not the abnormal point exists in the map to be checked. If the map to be checked has no abnormal points, executing step 308; if the map to be checked has an outlier, step 309 is performed.

And step 308, determining that the map to be audited is audited.

In this embodiment, if the map to be checked has no abnormal point, it may be determined that the map to be checked passes the check.

And for the map which passes the examination, the map can be added into a truth map library, so that the truth map library is expanded, and data is accumulated for subsequent automatic examination work.

Step 309, determining that the map review to be reviewed is not passed.

In this embodiment, if the map to be checked has an abnormal point, it may be determined that the map to be checked is not checked. For maps that do not pass the audit, step 210 continues.

Step 310, clustering the abnormal points in the abnormal point detection result to obtain an abnormal region of the map to be checked.

In this embodiment, at least one outlier cluster may be obtained for outlier clustering. The area where the abnormal point cluster is located is the abnormal area of the map to be checked.

And marking the abnormal region of the map to be audited, and outputting the marked abnormal region to an audit sending unit for restoration and correction. After repair and correction by the delivery unit, the map auditing method provided by the embodiment of the disclosure can be executed again to carry out next round of auditing.

As can be seen from fig. 3, compared with the embodiment corresponding to fig. 2, the map auditing method in this embodiment highlights the feature extraction step based on multi-scale line convolution, surface convolution, and the outlier detection step based on the adaptive feature matching algorithm. Therefore, the scheme described in the embodiment constructs an image feature set based on the multi-scale line convolution and the surface convolution feature extraction method, and detects abnormal points based on the self-adaptive feature matching algorithm, so that automatic aesthetic drawing is realized.

After the to-be-checked map is obtained from the checking unit, firstly, a truth value map is built based on a truth value library, then, an image feature set of the to-be-checked map and an image feature set of the truth value map are respectively built through a feature extraction algorithm based on multi-scale line convolution and surface convolution, and then, whether abnormal points exist is judged through a self-adaptive feature matching algorithm. If yes, the problem map is positioned through abnormal point clustering, and a review unit is returned for repair and correction; if not, the verification is passed, and the correct map is supplemented into a truth library, so that data is accumulated for subsequent automatic picture examination work.

FIG. 8 illustrates a scenario diagram of a map auditing method in which embodiments of the present disclosure may be implemented.

As shown in fig. 8, the overall flow of the automated image-examining algorithm based on image feature set similarity matching mainly includes:

1. map acquisition: and acquiring the review map based on a review machine provided by the review unit for automatic review.

2. Map extraction: and automatically intercepting the map to be checked for checking the map through an automatic screenshot tool and map element coordinate points based on the map to be checked.

3. True value map extraction: and constructing a truth value map based on the version and style of the historical review map for automatic review.

4. Constructing an image feature library: and extracting image features of multiple dimensions such as lines, planes and the like based on a feature extraction algorithm of multi-scale line and plane convolution, and constructing an image feature library.

5. Map outlier detection: based on the self-adaptive feature matching algorithm, comparing the image feature library of the map to be checked with the image feature library of the truth value map, automatically detecting abnormal points of the map, and outputting a suspected set of abnormal points of the problem map of the map to be checked.

In addition, the correct map without abnormal points of the map can be supplemented into a truth library, and data is accumulated for subsequent automatic picture examination work.

6. Clustering map outliers: and (3) checking out the picture points to be checked with abnormal points, and extracting a map abnormal region in a coordinate clustering mode for repairing and positioning by a checking unit.

7. And (5) outputting an abnormal area of the map to be examined: based on the result of automatic examination, the abnormal region is extracted and output to the examination unit for repair and correction.

With further reference to fig. 9, as an implementation of the method shown in the foregoing figures, the present disclosure provides an embodiment of a map auditing apparatus, where the apparatus embodiment corresponds to the method embodiment shown in fig. 2, and the apparatus may be specifically applied to various electronic devices.

As shown in fig. 9, the map auditing apparatus 900 of the present embodiment may include: an acquisition module 901, a selection module 902, a construction module 903 and a comparison module 904. Wherein, the acquisition module 901 is configured to acquire a map to be checked; a selection module 902 configured to select a truth value map corresponding to the map to be checked; the construction module 903 is configured to construct an image feature set of the map to be checked and an image feature set of the truth value map based on a feature extraction algorithm of the multi-scale line convolution and the face convolution; the comparison module 904 is configured to compare the image feature set of the map to be checked with the image feature set of the truth value map to obtain a checking result of the map to be checked.

In the present embodiment, in the map auditing apparatus 900: the specific processes and technical effects of the obtaining module 901, the selecting module 902, the constructing module 903 and the comparing module 904 may refer to the relevant descriptions of steps 201 to 204 in the corresponding embodiment of fig. 2, and are not described herein again.

In some alternative implementations of the present embodiment, the building block 903 includes: the line convolution sub-module is configured to conduct multi-scale line convolution on the map to be checked and the truth value map respectively to obtain a first image feature subset of the map to be checked and a first image feature subset of the truth value map; the surface convolution sub-module is configured to conduct multi-scale surface convolution on the map to be checked and the truth value map respectively to obtain a second image feature subset of the map to be checked and a second image feature subset of the truth value map; the merging sub-module is configured to merge the first image feature subset and the second image feature subset of the map to be checked to obtain an image feature set of the map to be checked, and merge the first image feature subset and the second image feature subset of the truth value map to obtain the image feature set of the truth value map.

In some optional implementations of the present embodiment, the wire convolution sub-module is further configured to: and carrying out multi-scale line convolution on the map to be audited and the truth value map based on at least one convolution kernel size and at least one deflection angle to obtain a first image feature subset of the map to be audited and a first image feature subset of the truth value map, wherein the convolution kernel size and the deflection angle are preset according to the convolution times of the multi-scale line convolution.

In some alternative implementations of the present embodiment, the face convolution sub-module is further configured to: slicing the feature map of the map to be checked and the feature map of the truth value map respectively according to the height direction and the width direction to obtain a feature slice map of the map to be checked and a feature slice map of the truth value map; and respectively migrating and fusing the features in the feature slice diagram of the map to be checked and the features in the feature slice diagram of the truth value map to a preset direction according to at least one migration span to obtain a second image feature subset of the map to be checked and a second image feature subset of the truth value map, wherein the migration span is preset according to the convolution times of the multi-scale surface convolution.

In some alternative implementations of the present embodiment, the comparison module 904 includes: and the matching sub-module is configured to compare the image feature set of the map to be checked with the image feature set of the truth value map by utilizing the self-adaptive feature matching network to obtain an abnormal point detection result of the map to be checked.

In some optional implementations of this embodiment, the matching submodule includes: the conversion unit is configured to respectively convert the features in the image feature set of the map to be checked and the features in the image feature set of the truth value map into feature vectors by using a convolutional neural network and/or a converter model in the adaptive feature matching network; the matching unit is configured to match the feature vector of the map to be checked with the corresponding feature vector of the truth value map to obtain an abnormal point detection result of the map to be checked.

In some optional implementations of the present embodiment, the matching unit is further configured to: calculating the similarity between the feature vector of the map to be checked and the corresponding feature vector of the truth value map; and carrying out coordinate reduction on the feature vectors with the similarity lower than the similarity preset threshold value to obtain abnormal points of the map to be checked.

In some optional implementations of the present embodiment, the matching unit is further configured to: calculating the difference value between the feature vector of the map to be checked and the feature vector of the surrounding area, and the difference value between the corresponding feature vector of the truth value map and the feature vector of the surrounding area, and comparing the difference of the two difference values; and carrying out coordinate reduction on the feature vectors with the difference larger than the difference preset threshold value to obtain abnormal points of the map to be checked.

In some optional implementations of the present embodiment, the map auditing apparatus 900 further includes: and the clustering module is configured to cluster the abnormal points in the abnormal point detection result to obtain an abnormal region of the map to be checked.

In some alternative implementations of the present embodiment, the selection module 902 is further configured to: selecting a truth value map consistent with the style of the map to be checked from a truth value map library, wherein the style comprises at least one of the following: typesetting, color, hue, and region.

In some optional implementations of the present embodiment, the map auditing apparatus 900 further includes: and the adding module is configured to add the checked map to be checked which passes the check into the truth value map library.

In the technical scheme of the disclosure, the acquisition, storage, application and the like of the related user personal information all conform to the regulations of related laws and regulations, and the public sequence is not violated.

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

Fig. 10 shows a schematic block diagram of an example electronic device 1000 that may 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. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular telephones, smartphones, 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. 10, the apparatus 1000 includes a computing unit 1001 that can perform various appropriate actions and processes according to a computer program stored in a Read Only Memory (ROM) 1002 or a computer program loaded from a storage unit 1008 into a Random Access Memory (RAM) 1003. In the RAM 1003, various programs and data required for the operation of the device 1000 can also be stored. The computing unit 1001, the ROM 1002, and the RAM 1003 are connected to each other by a bus 1004. An input/output (I/O) interface 1005 is also connected to bus 1004.

Various components in device 1000 are connected to I/O interface 1005, including: an input unit 1006 such as a keyboard, a mouse, and the like; an output unit 1007 such as various types of displays, speakers, and the like; a storage unit 1008 such as a magnetic disk, an optical disk, or the like; and communication unit 1009 such as a network card, modem, wireless communication transceiver, etc. Communication unit 1009 allows device 1000 to exchange information/data with other devices via a computer network, such as the internet, and/or various telecommunications networks.

The computing unit 1001 may be a variety of general and/or special purpose processing components having processing and computing capabilities. Some examples of computing unit 1001 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various specialized Artificial Intelligence (AI) computing chips, various computing units running machine learning model algorithms, a Digital Signal Processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 1001 performs the respective methods and processes described above, such as a map auditing method. For example, in some embodiments, the map auditing method may be implemented as a computer software program that is tangibly embodied on a machine-readable medium, such as the storage unit 1008. In some embodiments, part or all of the computer program may be loaded and/or installed onto device 1000 via ROM 1002 and/or communication unit 1009. When the computer program is loaded into RAM 1003 and executed by computing unit 1001, one or more steps of the map auditing method described above may be performed. Alternatively, in other embodiments, the computing unit 1001 may be configured to perform the map auditing method in any other suitable manner (e.g., by means of firmware).

Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuit systems, field Programmable Gate Arrays (FPGAs), application Specific Integrated Circuits (ASICs), application Specific Standard Products (ASSPs), systems On Chip (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs, the one or more computer programs may be executed and/or interpreted on a programmable system including at least one programmable processor, which may be a special purpose or general-purpose programmable processor, that may receive data and instructions from, and transmit data and instructions to, a storage system, at least one input device, and at least one output device.

Program code for carrying out methods of the present disclosure may be written in any combination of one or more programming languages. These program code 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 code, when executed by the processor or controller, causes the functions/operations specified in the flowchart and/or block diagram to be implemented. 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. The 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 portable 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 pointing device (e.g., a mouse or 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 may 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 input, speech input, or tactile input.

The systems and techniques described here can be implemented in a computing system that includes a background 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 background, 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 a client and a server. The client and server are typically 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 may be a cloud server, a server of a distributed system, or a server incorporating a blockchain.

It should be appreciated that various forms of the flows shown above may be used to reorder, add, or delete steps. For example, the steps recited in the present disclosure may be performed in parallel, sequentially, or in a different order, provided that the desired results of the technical solutions provided by the present disclosure are achieved, and are not limited herein.

The above detailed description should not be taken as limiting the scope of the present disclosure. It will be apparent to those skilled in the art that various modifications, combinations, sub-combinations and alternatives are possible, depending on design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present disclosure are intended to be included within the scope of the present disclosure.

Claims

1. A map auditing method, comprising:

acquiring a map to be checked;

selecting a true value map corresponding to the map to be checked;

Performing multi-scale line convolution on the map to be checked and the truth value map based on at least one convolution kernel size and at least one deflection angle to obtain a first image feature subset of the map to be checked and a first image feature subset of the truth value map, wherein the convolution kernel size and the deflection angle are preset according to the convolution times of the multi-scale line convolution; slicing the feature map of the map to be checked and the feature map of the truth value map respectively according to the height direction and the width direction to obtain a feature slice map of the map to be checked and a feature slice map of the truth value map; respectively migrating and fusing the features in the feature slice of the map to be checked and the features in the feature slice of the truth value map in sequence along a preset direction according to at least one migration span to obtain a second image feature subset of the map to be checked and a second image feature subset of the truth value map, wherein the migration span is preset according to the convolution times of multi-scale surface convolution; combining the first image feature subset and the second image feature subset of the map to be checked to obtain an image feature set of the map to be checked, and combining the first image feature subset and the second image feature subset of the truth value map to obtain an image feature set of the truth value map;

And comparing the image feature set of the map to be checked with the image feature set of the truth value map to obtain a checking result of the map to be checked.

2. The method of claim 1, wherein the comparing the image feature set of the map to be inspected with the image feature set of the truth map to obtain the inspection result of the map to be inspected comprises:

and comparing the image feature set of the map to be checked with the image feature set of the truth value map by utilizing a self-adaptive feature matching network to obtain an abnormal point detection result of the map to be checked.

3. The method of claim 2, wherein the comparing the image feature set of the map to be inspected with the image feature set of the truth map using the adaptive feature matching network to obtain the outlier detection result of the map to be inspected comprises:

respectively converting the features in the image feature set of the map to be checked and the features in the image feature set of the truth value map into feature vectors by using a convolutional neural network and/or a converter model in the self-adaptive feature matching network;

And matching the feature vector of the map to be checked with the corresponding feature vector of the truth value map to obtain an abnormal point detection result of the map to be checked.

4. The method of claim 3, wherein the matching the feature vector of the to-be-inspected map with the corresponding feature vector of the truth value map to obtain the outlier detection result of the to-be-inspected map includes:

calculating the similarity between the feature vector of the map to be checked and the corresponding feature vector of the truth value map;

And carrying out coordinate reduction on the feature vectors with the similarity lower than the similarity preset threshold value to obtain the abnormal points of the map to be checked.

5. The method according to claim 3 or 4, wherein the matching the feature vector of the map to be checked with the corresponding feature vector of the truth value map to obtain the outlier detection result of the map to be checked includes:

Calculating the difference value between the feature vector of the map to be checked and the feature vector of the surrounding area, and the difference value between the corresponding feature vector of the truth value map and the feature vector of the surrounding area, and comparing the difference value between the two difference values;

And carrying out coordinate reduction on the feature vectors with the difference larger than a preset difference threshold value to obtain the abnormal points of the map to be checked.

6. The method of claim 2, wherein the method further comprises:

Clustering the abnormal points in the abnormal point detection result to obtain the abnormal region of the map to be checked.

7. The method of claim 1, wherein the selecting the truth value map corresponding to the map under review comprises:

Selecting a truth value map consistent with the style of the map to be checked from a truth value map library, wherein the style comprises at least one of the following: typesetting, color, hue, and region.

8. The method of claim 7, wherein the method further comprises:

And adding the map to be checked which passes the checking into the truth value map library.

9. A map auditing apparatus, comprising:

the acquisition module is configured to acquire the map to be checked;

the selection module is configured to select a true value map corresponding to the map to be checked;

the construction module is configured to respectively carry out multi-scale line convolution on the map to be checked and the truth value map based on at least one convolution kernel size and at least one deflection angle to obtain a first image feature subset of the map to be checked and a first image feature subset of the truth value map, wherein the convolution kernel size and the deflection angle are preset according to the convolution times of the multi-scale line convolution; slicing the feature map of the map to be checked and the feature map of the truth value map respectively according to the height direction and the width direction to obtain a feature slice map of the map to be checked and a feature slice map of the truth value map; respectively migrating and fusing the features in the feature slice of the map to be checked and the features in the feature slice of the truth value map in sequence along a preset direction according to at least one migration span to obtain a second image feature subset of the map to be checked and a second image feature subset of the truth value map, wherein the migration span is preset according to the convolution times of multi-scale surface convolution; combining the first image feature subset and the second image feature subset of the map to be checked to obtain an image feature set of the map to be checked, and combining the first image feature subset and the second image feature subset of the truth value map to obtain an image feature set of the truth value map;

And the comparison module is configured to compare the image feature set of the map to be checked with the image feature set of the truth value map to obtain a checking result of the map to be checked.

10. The apparatus of claim 9, wherein the comparison module comprises:

And the matching sub-module is configured to compare the image feature set of the map to be checked with the image feature set of the truth value map by utilizing the self-adaptive feature matching network to obtain an abnormal point detection result of the map to be checked.

11. The apparatus of claim 10, wherein the matching submodule comprises:

A conversion unit configured to convert features in the image feature set of the map to be checked and features in the image feature set of the truth value map into feature vectors respectively using a convolutional neural network and/or a converter model in the adaptive feature matching network;

and the matching unit is configured to match the feature vector of the map to be checked with the corresponding feature vector of the truth value map to obtain an abnormal point detection result of the map to be checked.

12. The apparatus of claim 11, wherein the matching unit is further configured to:

13. The apparatus of claim 11 or 12, wherein the matching unit is further configured to:

14. The apparatus of claim 10, wherein the apparatus further comprises:

and the clustering module is configured to cluster the abnormal points in the abnormal point detection result to obtain the abnormal region of the map to be checked.

15. The apparatus of claim 9, wherein the selection module is further configured to:

16. The apparatus of claim 15, wherein the apparatus further comprises:

and the adding module is configured to add the map to be checked which passes the check into the truth value map library.

17. An electronic device, comprising:

at least one processor; and

A memory communicatively coupled to the at least one processor; wherein,

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 1-8.

18. A non-transitory computer readable storage medium storing computer instructions for causing the computer to perform the method of any one of claims 1-8.