CN114463755A - Automatic sensitive information detection desensitization method in high-precision map-based acquired picture - Google Patents
Automatic sensitive information detection desensitization method in high-precision map-based acquired picture Download PDFInfo
- Publication number
- CN114463755A CN114463755A CN202111527129.4A CN202111527129A CN114463755A CN 114463755 A CN114463755 A CN 114463755A CN 202111527129 A CN202111527129 A CN 202111527129A CN 114463755 A CN114463755 A CN 114463755A
- Authority
- CN
- China
- Prior art keywords
- character
- sensitive
- information
- picture
- signboards
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/60—Protecting data
- G06F21/62—Protecting access to data via a platform, e.g. using keys or access control rules
- G06F21/6218—Protecting access to data via a platform, e.g. using keys or access control rules to a system of files or objects, e.g. local or distributed file system or database
- G06F21/6245—Protecting personal data, e.g. for financial or medical purposes
- G06F21/6254—Protecting personal data, e.g. for financial or medical purposes by anonymising data, e.g. decorrelating personal data from the owner's identification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Databases & Information Systems (AREA)
- Bioethics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Medical Informatics (AREA)
- Computer Hardware Design (AREA)
- Computer Security & Cryptography (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to a method for automatically detecting and desensitizing sensitive information in a collected picture based on a high-precision map, which comprises the following steps: making a key sensitive character table, wherein the key sensitive character table stores each classified sensitive character content; extracting a character signboard in a series of picture frames collected by a collection vehicle, identifying characters in the character signboard, searching and matching an identification result with a key sensitive character table, determining whether the character signboard has sensitive character content according to a matching result, determining corresponding classification information when the character signboard is judged to have the sensitive character content, and performing local fuzzy processing on the picture frames of the signboard containing the sensitive character content; in the map making specification, because the sensitive character information is more and the related range of the sensitive character information is expanded, before the sensitive information is matched, a sensitive character information table is constructed, so that the matching can be assisted and quickly carried out, and the automatic detection and elimination operation of the key character information in the picture is completed.
Description
Technical Field
The invention relates to the field of high-precision map manufacturing, in particular to a method for automatically detecting and desensitizing sensitive information in a collected picture based on a high-precision map.
Background
In the high-precision map making process, in order to ensure that high-precision navigation electronic maps and drive test service acquisition data produced by companies meet the national standard, character signboards in the maps need to be detected, sensitive information is labeled, and the high-precision character signboards are extracted in advance, so that sensitive character and picture information cannot exist in high-precision map data.
When the traditional method is adopted for target extraction, the extraction speed is low, and the applicable scene is single.
Disclosure of Invention
Aiming at the technical problems in the prior art, the invention provides an automatic sensitive information detection desensitization method based on a high-precision map acquisition picture, aiming at the problem that sensitive information secret-related keywords in the picture acquired by an acquisition vehicle need to be extracted and removed for picture desensitization; in the map making specification, because the sensitive character information is more and the related range of the sensitive character information is expanded, the sensitive character information table is constructed before the sensitive information is matched, and the map making specification can assist in quickly matching.
According to the first aspect of the invention, a method for automatically detecting desensitization based on sensitive information in a high-precision map acquisition picture is provided, which comprises the following steps: making a key sensitive character table, wherein the classified sensitive character contents are stored in the key sensitive character table;
extracting the character signboards in the series of picture frames collected by the collection vehicle, identifying characters in the character signboards, searching and matching the identification results with the key sensitive character table, determining whether the character signboards have sensitive character contents according to the matching results, determining corresponding classification information when the sensitive character contents are determined, and performing local fuzzy processing on the picture frames of the signboards containing the sensitive character contents.
On the basis of the technical scheme, the invention can be improved as follows.
Optionally, the classifying the content of each inscription in the key-sensitive character table includes:
the classification according to the content information comprises the following steps: important geographic locations, confidential units, and public safety facilities;
classifying according to the sensitive type of the content, comprising: sensitive keywords and semi-fixed sensitive words.
Optionally, the making the key sensitive text table further includes: and constructing indexes of the classification labels and the sensitive text content.
Optionally, the extracting the text signboard in the series of picture frames collected by the collection vehicle includes:
acquiring a series of picture frames acquired by a high-precision map acquisition vehicle;
and obtaining a target detection model of the character signboard based on the training in the series of picture frames, extracting the character signboard based on the trained target detection model, and cutting a character target area in the character signboard.
Optionally, the target detection model adopts a multi-level transform visual model, performs target detection based on a skeleton framework of a self-attention mechanism, extracts geometric information in the text signboard in the picture frame through the bounding box, and acquires picture content information cut in the bounding box of the text signboard.
Optionally, after the text signboard is cut, identifying the text in the text signboard includes:
adopting a text detection network based on segmentation, acquiring a binary mask image of the cut character signboard through a fixed threshold, extracting a contour in the mask image to acquire position information of a sentence character, and acquiring picture content information corresponding to the position information;
and performing character recognition on the extracted sentence characters by adopting an RNN (radio network) for sequence prediction, extracting the characteristics of the detected sentence character small pictures through a convolutional neural network, predicting the sequence by adopting a cyclic neural network, and obtaining a final character recognition result through a translation layer of a CTC (China traffic control) algorithm.
Optionally, after the local blurring processing is performed on the picture frame of the signboard containing the sensitive text content, the method further includes:
feeding back the geometric information of the operated character signboards of the series of picture frames and the corresponding classification information to a high-precision map making system, and replacing the original picture by using the picture frames after local blurring.
According to a second aspect of the present invention, there is provided a system for automatically detecting desensitization based on sensitive information in a high-precision map captured picture, comprising: the key sensitive character list making module and the fuzzy processing module;
the key sensitive character table making module is used for making a key sensitive character table, and the key sensitive character table stores the classified sensitive character contents;
the fuzzy processing module is used for extracting the character signboards in the series of picture frames collected by the collection vehicle, after characters in the character signboards are identified, searching and matching the identification results with the key sensitive character table, determining whether the character signboards have sensitive character contents according to the matching results, determining corresponding classification information when the sensitive character contents are judged, and performing local fuzzy processing on the picture frames of the signboards containing the sensitive character contents.
According to a third aspect of the present invention, there is provided an electronic device comprising a memory, a processor for implementing the steps of the method for automatically detecting desensitization based on sensitive information in high-precision map-captured pictures when executing a computer management-like program stored in the memory.
According to a fourth aspect of the present invention, there is provided a computer readable storage medium having stored thereon a computer management like program, which when executed by a processor, performs the steps of automatically detecting a desensitization method based on sensitive information in a high-precision map-captured picture.
According to the automatic sensitive information detection desensitization method, the system, the electronic equipment and the storage medium in the high-precision map-based acquired pictures, the character signboards in the sequence frame pictures are detected and then are subjected to character detection and identification, a deep learning model is simple to use, the detection and identification can be rapidly carried out, the consumed time is short, and the sublimation ability is strong; meanwhile, the matching operation of the sensitive character information can be accelerated by tabulating the sensitive character information; sensitive character information in the signboard is detected and identified, simple deep learning model detection of CNN + RNN + CTC is mainly applied, the model is simple and short in time consumption, and sensitive characters in the character signboard can be rapidly and accurately identified; the local small sensitive character area is fuzzified to replace the original frame picture, so that the integrity of the collected picture information is kept as much as possible, and the operation is simple, convenient and quick.
Drawings
Fig. 1 is a flowchart of automatic detection desensitization based on sensitive information in a high-precision map acquired picture according to an embodiment of the present invention;
fig. 2(a) is a schematic diagram of a single picture in a series of picture frames acquired by a high-precision map acquisition vehicle according to an embodiment of the present invention;
fig. 2(b) is a diagram of the effect of the text sign extraction and the text detection and recognition in the sign provided by the embodiment of the present invention;
fig. 3(a) is a schematic diagram of a picture frame acquired by an acquisition vehicle according to an embodiment of the present invention;
fig. 3(b) is a schematic diagram of text information extracted from a picture according to an embodiment of the present invention;
fig. 3(c) is a schematic diagram illustrating an effect after local small region blurring processing is performed according to an embodiment of the present invention;
FIG. 4 is a structural diagram of automatic detection desensitization based on sensitive information in a high-precision map acquisition picture according to the present invention;
FIG. 5 is a schematic diagram of a hardware structure of a possible electronic device provided in the present invention;
fig. 6 is a schematic diagram of a hardware structure of a possible computer-readable storage medium according to the present invention.
Detailed Description
The principles and features of this invention are described below in conjunction with the following drawings, which are set forth by way of illustration only and are not intended to limit the scope of the invention.
Fig. 1 is a flowchart of an automatic detection desensitization method based on sensitive information in a high-precision map acquired picture according to an embodiment of the present invention, and as shown in fig. 1, the automatic detection desensitization method includes:
and making a key sensitive character table, wherein the classified sensitive character contents are stored in the key sensitive character table.
Extracting the character signboards in the series of picture frames collected by the collecting vehicle, identifying characters in the character signboards, searching and matching the identification results with the key sensitive character table, determining whether the character signboards have sensitive character contents according to the matching results, determining corresponding classification information when the character signboards have the sensitive character contents, and performing local fuzzy processing on the picture frames of the signboards containing the sensitive character contents.
Aiming at the problem that sensitive information secret-related keywords in a picture acquired by an acquisition vehicle need to be extracted and removed for picture desensitization, the invention provides a method for automatically detecting and desensitizing sensitive information in a map acquisition picture based on a high-precision map, which is used for detecting key characters in the picture acquired by the map acquisition vehicle and completing the operation of automatically detecting and removing the key character information in the picture so as to automatically extract and desensitize character information of the acquired picture; in the map making specification, because the sensitive character information is more and the related range of the sensitive character information is expanded, the sensitive character information table is constructed before the sensitive information is matched, and the map making specification can assist in quickly matching.
Example 1
Embodiment 1 provided by the present invention is an embodiment of an automatic detection desensitization method based on sensitive information in a high-precision map-acquired picture provided by the present invention, and as can be seen from fig. 2, the embodiment of the automatic detection desensitization method includes:
and making a key sensitive character table, wherein the classified sensitive character contents are stored in the key sensitive character table.
In one possible embodiment, the classifying the respective inscription contents in the key-sensitive character table includes:
the classification according to the content information comprises the following steps: important geographic locations, confidential units, and public safety facilities.
Classifying according to the sensitive type of the content, comprising: sensitive keywords and semi-fixed sensitive words.
In one possible embodiment, creating the key-sensitive text table further comprises: and constructing indexes of the classification labels and the sensitive text content.
In specific implementation, the key sensitive character table can be manually determined after statistics is carried out according to character signboards collected by a large number of collection vehicles. The required key sensitive words are classified, and content information is classified according to the sensitive contents of the key sensitive words, wherein the classification and tabulation comprise important geographic positions, confidential units, public safety facilities and the like. And meanwhile, continuously dividing fixed sensitive keywords (such as water supply) and semi-fixed sensitive words (such as a certain power station) in the classification tabulation to construct a label-sensitive word content index.
Extracting the character signboards in the series of picture frames collected by the collecting vehicle, identifying characters in the character signboards, searching and matching the identification results with the key sensitive character table, determining whether the character signboards have sensitive character contents according to the matching results, determining corresponding classification information when the character signboards have the sensitive character contents, and performing local fuzzy processing on the picture frames of the signboards containing the sensitive character contents.
In one possible embodiment, the extracting the text signboard in the series of picture frames acquired by the acquisition vehicle comprises:
and acquiring a series of picture frames acquired by the high-precision map acquisition vehicle.
And training based on the series of picture frames to obtain a target detection model of the character signboard, extracting the character signboard based on the trained target detection model, and cutting a character target area in the character signboard.
In a possible embodiment mode, the target detection model adopts a multi-level transform visual model, performs target detection based on a skeleton framework of a self-attention mechanism, extracts geometric information in a character signboard in a picture frame through a bounding box, and acquires picture content information cut in the bounding box of the character signboard.
The deep learning target detection model is used for detecting the character signboards, the speed is high, the model generalization capability is strong, and the method can be suitable for detecting scenes of various road character signboards.
In one possible embodiment, after cutting the text signboard, the recognizing the text in the text signboard comprises:
and acquiring a binary mask image of the cut character signboard through a fixed threshold value by adopting a text detection network based on segmentation, extracting the outline in the mask image to acquire the position information of the sentence character, and acquiring the picture content information corresponding to the position information.
The extracted sentence characters are subjected to character recognition by adopting an RNN (radio network) for predicting sequences, the characteristics of the detected sentence character small pictures are extracted by adopting a convolutional neural network, the sequences are predicted by adopting a cyclic neural network, and a final character recognition result is obtained by a translation layer of a CTC (continuity test Classification) algorithm.
The sensitive character information in the signboard is detected and identified by mainly applying simple deep learning model detection of CNN + RNN + CTC, and the model is simple, short in time consumption and capable of quickly and accurately identifying the sensitive characters in the character signboard. Then, the local small area of the sensitive characters is processed in a fuzzy mode, so that original picture information is reserved, and the operation is simple, convenient and fast.
In a possible embodiment mode, after the local blurring processing is performed on the picture frame of the signboard containing the sensitive text content, the method further comprises the following steps:
and feeding back the geometric information of the character signboards of the operated serial picture frames and the corresponding classification information to a high-precision map making system, and replacing the original picture by using the picture frames after local blurring to finish automatic detection and desensitization of sensitive character information.
Fig. 2 is a diagram illustrating extraction of a character signboard from a picture collected by a collection vehicle and detection and recognition of a corresponding character in the picture, fig. 2(a) is a schematic diagram illustrating a single-frame picture in a series of picture frames collected by a high-precision map collection vehicle according to an embodiment of the present invention, and fig. 2(b) is an effect diagram after extraction of the character signboard and detection and recognition of a character in the signboard according to an embodiment of the present invention.
Fig. 3 is a diagram of a fuzzy operation of capturing vehicle sensitive characters, and fig. 3(a) is a schematic diagram of a picture frame captured by a capturing vehicle according to an embodiment of the present invention; fig. 3(b) is a schematic diagram of text information extracted from a picture according to an embodiment of the present invention; fig. 3(c) is a schematic diagram illustrating an effect after local small region blurring processing is performed according to an embodiment of the present invention; in fig. 3(c), the "dangerous goods vehicle" in the picture is taken as an example, and if the "dangerous goods vehicle" is in the sensitive character information table, the local small region blurring processing is performed.
Example 2
Embodiment 2 of the present invention is an embodiment of an automatic detection desensitization system for sensitive information in a high-precision map-based acquired picture according to the present invention, and fig. 4 is a structural diagram of an automatic detection desensitization system for sensitive information in a high-precision map-based acquired picture according to an embodiment of the present invention, and it can be seen from fig. 4 that the embodiment of the automatic detection desensitization system includes: a key sensitive character table making module and a fuzzy processing module.
And the key sensitive character table manufacturing module is used for manufacturing a key sensitive character table, and the classified sensitive character contents are stored in the key sensitive character table.
And the fuzzy processing module is used for extracting the character signboards in the series of picture frames acquired by the acquisition vehicle, identifying characters in the character signboards, searching and matching the identification results with the key sensitive character table, determining whether the character signboards have sensitive character contents according to the matching results, determining corresponding classification information when the sensitive character contents are determined, and performing local fuzzy processing on the picture frames of the signboards containing the sensitive character contents.
It can be understood that the automatic detection desensitization system for sensitive information in the collected picture based on the high-precision map provided by the invention corresponds to the automatic detection desensitization method for sensitive information in the collected picture based on the high-precision map provided by the embodiments, and the relevant technical features of the automatic detection desensitization system for sensitive information in the collected picture based on the high-precision map can refer to the relevant technical features of the automatic detection desensitization method for sensitive information in the collected picture based on the high-precision map, and are not described herein again.
Referring to fig. 5, fig. 5 is a schematic view of an embodiment of an electronic device according to an embodiment of the invention. As shown in fig. 5, an embodiment of the present invention provides an electronic device, which includes a memory 1310, a processor 1320, and a computer program 1311 stored in the memory 1310 and executable on the processor 1320, where the processor 1320 executes the computer program 1311 to implement the following steps: making a key sensitive character table, wherein the key sensitive character table stores each classified sensitive character content; extracting the character signboards in the series of picture frames collected by the collecting vehicle, identifying characters in the character signboards, searching and matching the identification results with the key sensitive character table, determining whether the character signboards have sensitive character contents according to the matching results, determining corresponding classification information when the sensitive character contents are determined, and performing local fuzzy processing on the picture frames of the signboards containing the sensitive character contents.
Referring to fig. 6, fig. 6 is a schematic diagram of an embodiment of a computer-readable storage medium according to the present invention. As shown in fig. 6, the present embodiment provides a computer-readable storage medium 1400, on which a computer program 1411 is stored, which computer program 1411, when executed by a processor, implements the steps of: making a key sensitive character table, wherein the key sensitive character table stores each classified sensitive character content; extracting the character signboards in the series of picture frames collected by the collecting vehicle, identifying characters in the character signboards, searching and matching the identification results with the key sensitive character table, determining whether the character signboards have sensitive character contents according to the matching results, determining corresponding classification information when the character signboards have the sensitive character contents, and performing local fuzzy processing on the picture frames of the signboards containing the sensitive character contents.
The embodiment of the invention provides a method, a system, electronic equipment and a storage medium for automatic detection desensitization of sensitive information in a high-precision map-collected picture, and a transformer target detection method is adopted to predict and acquire a bounding box of a character signboard. And then carrying out character detection and identification on the cut character signboard area, then carrying out matching search with a table made of sensitive characters, obtaining the character signboard containing sensitive character information, carrying out sensitive character information classification marking and local sensitive character small area blurring, and finishing the automatic detection desensitization operation of the sensitive information in the high-precision map acquisition picture. The deep learning target detection model is used for detecting the character signboards, the speed is high, the model generalization capability is strong, and the method can be suitable for detecting scenes of various road character signboards. The sensitive character information in the signboard is detected and identified by mainly applying simple deep learning model detection of CNN + RNN + CTC, the model is simple and short in time consumption, and the sensitive characters in the character signboard can be quickly and accurately identified. The local small area of the sensitive characters is processed in a fuzzy mode, so that original picture information is reserved, and the operation is simple, convenient and fast.
It should be noted that, in the foregoing embodiments, the descriptions of the respective embodiments have respective emphasis, and reference may be made to relevant descriptions of other embodiments for parts that are not described in detail in a certain embodiment.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
While preferred embodiments of the present invention have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. Therefore, it is intended that the appended claims be interpreted as including preferred embodiments and all such alterations and modifications as fall within the scope of the invention.
It will be apparent to those skilled in the art that various changes and modifications may be made in the present invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.
Claims (10)
1. A sensitive information automatic detection desensitization method in a collected picture based on a high-precision map is characterized by comprising the following steps:
making a key sensitive character table, wherein the classified sensitive character contents are stored in the key sensitive character table;
extracting the character signboards in the series of picture frames collected by the collection vehicle, identifying characters in the character signboards, searching and matching the identification results with the key sensitive character table, determining whether the character signboards have sensitive character contents according to the matching results, determining corresponding classification information when the sensitive character contents are determined, and performing local fuzzy processing on the picture frames of the signboards containing the sensitive character contents.
2. The automatic detection desensitization method of claim 1, wherein classifying individual nameplate text in said key sensitive text table comprises:
the classification according to the content information comprises the following steps: important geographic locations, confidential units, and public safety facilities;
classifying according to the sensitive type of the content, comprising: sensitive keywords and semi-fixed sensitive words.
3. The method of automatic detection desensitization of claim 1, wherein said producing a key sensitive text table further comprises: and constructing indexes of the classification labels and the sensitive text content.
4. The automatic detection desensitization method according to claim 1, wherein said extracting the textual signs in the series of picture frames captured by the capture cart comprises:
acquiring a series of picture frames acquired by a high-precision map acquisition vehicle;
and training to obtain a target detection model of the character signboard based on the series of picture frames, extracting the character signboard based on the trained target detection model, and cutting a character target area in the character signboard.
5. The automatic detection desensitization method according to claim 4, wherein the target detection model employs a multi-level transform visual model, performs target detection based on a skeleton structure of a self-attention mechanism, extracts geometric information in a text signboard in a picture frame through a bounding box, and obtains picture content information clipped in the bounding box of the text signboard.
6. The automatic detection desensitization method of claim 4, wherein identifying the text in the text label after cutting the text label comprises:
adopting a text detection network based on segmentation, acquiring a binary mask image of the cut character signboard through a fixed threshold, extracting a contour in the mask image to acquire position information of a sentence character, and acquiring picture content information corresponding to the position information;
and performing character recognition on the extracted sentence characters by adopting an RNN (radio network) for sequence prediction, extracting the characteristics of the detected sentence character small picture through a convolutional neural network, predicting the sequence by adopting a cyclic neural network, and obtaining a final character recognition result through a translation layer of a CTC (central control point) algorithm.
7. The automatic detection desensitization method according to claim 6, wherein said local blurring of the picture frames of the signboards containing sensitive text further comprises:
feeding back the geometric information of the operated character signboards of the series of picture frames and the corresponding classification information to a high-precision map making system, and replacing the original picture by using the picture frames after local blurring.
8. An automatic detection desensitization system based on sensitive information in a high-precision map acquisition picture is characterized by comprising: the key sensitive character list making module and the fuzzy processing module;
the key sensitive character table making module is used for making a key sensitive character table, and the key sensitive character table stores the classified sensitive character contents;
the fuzzy processing module is used for extracting the character signboards in the series of picture frames collected by the collection vehicle, after characters in the character signboards are identified, searching and matching the identification results with the key sensitive character table, determining whether the character signboards have sensitive character contents according to the matching results, determining corresponding classification information when the sensitive character contents are judged, and performing local fuzzy processing on the picture frames of the signboards containing the sensitive character contents.
9. An electronic device, comprising a memory and a processor, wherein the processor is used for implementing the steps of the method for automatic detection of desensitization based on sensitive information in high-precision map acquisition pictures according to any one of claims 1 to 7 when executing a computer management program stored in the memory.
10. A computer-readable storage medium, having stored thereon a computer management-like program, which when executed by a processor, performs the steps of the method for automatically detecting desensitization based on sensitive information in high-precision map-captured pictures according to any of claims 1-7.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111527129.4A CN114463755A (en) | 2021-12-13 | 2021-12-13 | Automatic sensitive information detection desensitization method in high-precision map-based acquired picture |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111527129.4A CN114463755A (en) | 2021-12-13 | 2021-12-13 | Automatic sensitive information detection desensitization method in high-precision map-based acquired picture |
Publications (1)
Publication Number | Publication Date |
---|---|
CN114463755A true CN114463755A (en) | 2022-05-10 |
Family
ID=81406724
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202111527129.4A Pending CN114463755A (en) | 2021-12-13 | 2021-12-13 | Automatic sensitive information detection desensitization method in high-precision map-based acquired picture |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114463755A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117455751A (en) * | 2023-12-22 | 2024-01-26 | 新华三网络信息安全软件有限公司 | Road section image processing system and method |
-
2021
- 2021-12-13 CN CN202111527129.4A patent/CN114463755A/en active Pending
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117455751A (en) * | 2023-12-22 | 2024-01-26 | 新华三网络信息安全软件有限公司 | Road section image processing system and method |
CN117455751B (en) * | 2023-12-22 | 2024-03-26 | 新华三网络信息安全软件有限公司 | Road section image processing system and method |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN113688652B (en) | Abnormal driving behavior processing method and device | |
CN102289686B (en) | Method for identifying classes of moving targets based on transfer learning | |
CN107992819B (en) | Method and device for determining vehicle attribute structural features | |
CN110781711A (en) | Target object identification method and device, electronic equipment and storage medium | |
CN111582126B (en) | Pedestrian re-recognition method based on multi-scale pedestrian contour segmentation fusion | |
Al Faqheri et al. | A real-time Malaysian automatic license plate recognition (M-ALPR) using hybrid fuzzy | |
Islam et al. | Automatic vehicle number plate recognition using structured elements | |
CN108960175A (en) | A kind of licence plate recognition method based on deep learning | |
CN111008574A (en) | Key person track analysis method based on body shape recognition technology | |
CN111178282A (en) | Road traffic speed limit sign positioning and identifying method and device | |
Dorbe et al. | FCN and LSTM based computer vision system for recognition of vehicle type, license plate number, and registration country | |
CN112733666A (en) | Method, equipment and storage medium for collecting difficult images and training models | |
Latha et al. | Image understanding: semantic segmentation of graphics and text using faster-RCNN | |
Gupta et al. | Detection of Number Plate in Vehicles using Deep Learning based Image Labeler Model | |
CN115272222A (en) | Method, device and equipment for processing road detection information and storage medium | |
CN114463755A (en) | Automatic sensitive information detection desensitization method in high-precision map-based acquired picture | |
CN114064971A (en) | Airport apron video semantic retrieval method and retrieval system based on deep learning | |
Arief et al. | Traffic signs detection and recognition system using the YOLOv4 algorithm | |
CN115131826B (en) | Article detection and identification method, and network model training method and device | |
CN116580232A (en) | Automatic image labeling method and system and electronic equipment | |
Vinay et al. | Automatic Number Plate Recognition for Different Fonts and Non-Roman Script | |
Rigoll et al. | Focus on the Challenges: Analysis of a User-friendly Data Search Approach with CLIP in the Automotive Domain | |
Das et al. | Object Detection on Scene Images: A Novel Approach | |
CN111259760A (en) | Dynamic target behavior identification method and device | |
Dhyani et al. | Real-Time License Plate Detection and Recognition System using YOLOv7x and EasyOCR |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |